1 ESSENTIAL TREMOR: PARADOXICAL VISUOSPATIAL COGNITION By UTAKA S. SPRINGER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORID A IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010
2 2010 Utaka S. Springer
3 To the volunteers of this study
4 ACKNOWLEDGMENTS First and foremost, I thank my graduate school mentor and dissertation committee chair, Dr. Dawn Bowers, for her guidance and encouragement. Thes e were invaluable to me throughout the study, but also thr oughout my formative years in academic and clinical neuropsychology. Many others provided me with valuable feedback during the proposal and early stages of this study including members of my dissertation committee (Drs. Russell Baue r, Hubert Fernandez, Yijun Liu, Michael Marsiske, Christina McCrae, and Robert Garrigues). Drs. William Perlstein and Keith White were helpful in providing guidance on RT and e rror rate data reduction and analysis. For their recruitment help, I thank Janet Romrell and Dr. Irene Malaty of the UF Movement Disorders Center, Shirley Bloodwo rth of the PrimeTime senior activity group at Santa Fe College, Sara Lynn McCrea of the Oak Hammo ck retirement community in Gainesville, FL, and the Villages / Lady Lake, FL essent ial tremor support group. Many students also lent their support during this projec t, including the members of Dr. BowersÂ’ Cognitive Neuroscience Laborat ory (Jenna Dietz, Lizabeth Jor dan, Daniel Kay, Lindsey Kirsch-Darrow, Ania Mikos, and Laura Zahod ne). Taylor Kuhn volunteered many hours during the early stages of this project, and Allie Clark prov ided helpful administrative support. Dr. BauerÂ’s graduate students Emily King and Gila Reckess were helpful in consultation about C-G Arena task parameters and the spatial navigation literature. All of these students are deeply thanked for graci ously lending me their time and insights. Finally, I thank my parents Marty and Michiko, to whom I give credit for inspiring my early interest in science, and my wife Shauna Springer, who has been a tremendous source of support and inspiration for me.
5 TABLE OF CONTENTS page ACKNOWLEDG MENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 8 LIST OF FI GURES .......................................................................................................... 9 ABSTRACT ................................................................................................................... 10 CHAPTER 1 OVERVIEW A ND AIMS .......................................................................................... 12 2 BACKGROUND AND S IGNIFICA NCE ................................................................... 15 Essential Trem or (ET) ............................................................................................. 15 Cognitive Functi oning in ET .................................................................................... 17 A Preliminary Study: Surprisi ng Visuospatial Findings ........................................... 18 Methodological Concerns ....................................................................................... 20 Â“SpatialÂ” vs. Â“Object/FormÂ” Processing: Subtypes of Visuospatial Cognition ......... 21 A Hypothesis for Superior Spat ial Cognition in ET .................................................. 23 Neuroplasticity and Vis uospatial C ognition ............................................................. 25 Tremor Severity vs. Tremor Duration: Potential Visuospati al Predictors................. 27 Spatial Cognition Subtypes: Fo ci of the Pres ent Study .......................................... 29 Mental Rota tion ................................................................................................ 30 Visual S earch ................................................................................................... 32 Spatial Memory and Navigation ........................................................................ 34 3 SUMMARY OF THE PROBLEM, SPECI FIC AIMS, AND H YPOTHESES .............. 37 Summary of t he Problem ........................................................................................ 37 Specific Aim 1 ......................................................................................................... 39 Overview .......................................................................................................... 39 Hypotheses ...................................................................................................... 39 Predictions ........................................................................................................ 40 Analyse s ........................................................................................................... 40 Specific Aim 2 ......................................................................................................... 40 Overview .......................................................................................................... 41 Hypothesis, Predicti on, Anal ysis ...................................................................... 41 4 MATERIALS A ND METHOD S ................................................................................ 42 Design Over view ..................................................................................................... 42 Partici pants ............................................................................................................. 42 Recruitment and Initial In clusion Crit eria .......................................................... 42
6 Screening and Exclus ion Criter ia ..................................................................... 44 Participant Char acterist ics ................................................................................ 46 Fronto-Executive F unctioning Te sts ....................................................................... 50 Overview and Ra tionale ................................................................................... 50 Tests of Working Memo ry / Atte ntion ............................................................... 50 Â“Digit SpanÂ” subtes t ................................................................................... 50 Â“Spatial SpanÂ” subtes t ................................................................................ 51 Test of Inhibition: Â“Stroop Color and Word Â” Test .............................................. 51 Tests of Verbal Fluency: Â“Letter Fl uencyÂ” & Â“Category Fl uencyÂ” Tests ............ 52 Visual Cognit ion Tests ............................................................................................ 52 Overview & Rational e ....................................................................................... 52 Clinical Neuropsychol ogical Te sts .................................................................... 53 Â“Judgment of Line Orientat ionÂ” (JOLO) test ............................................... 53 Â“Facial Recognition TestÂ” (FRT) ................................................................. 53 Computerized Experim ental Ta sks ................................................................... 54 Common physical set up across tasks ........................................................ 54 Â“Choice Reaction Ti meÂ” (CRT) task ........................................................... 57 Â“Mental Rotation TaskÂ” (MRT) .................................................................... 59 Â“Visual Search TaskÂ” (VST) ....................................................................... 63 Â“Computer-Generated ArenaÂ” (C-G Arena) task ........................................ 68 Â“Object Attention TaskÂ” (OAT) .................................................................... 73 Data Reduction & Analysis Plan for t he Stud y ........................................................ 79 5 RESULT S ............................................................................................................... 82 Specific Aim 1: Group Co mparisons on Â“SpatialÂ” and Â“O bject/FormÂ” Measures .... 82 Overview .......................................................................................................... 82 Demographic, Cognitive and Mood Vari ables ................................................. 82 CRT Resu lts ..................................................................................................... 85 Spatial Tasks (JOLO, MRT, VST, CG Arena): Group Co mparisons ............... 87 JOLO results .............................................................................................. 87 MRT results ................................................................................................ 89 VST resu lts ................................................................................................ 93 C-G Arena re sults ...................................................................................... 97 Object/Form Tasks (FRT, OAT) : Group Compar isons ................................... 100 FRT result s .............................................................................................. 100 OAT result s .............................................................................................. 102 Summary of Specific Aim Results: Group Comparisons on Spatial vs. Object/Form Tasks ...................................................................................... 105 Specific Aim 2: Controlling for Other Variables in the Spat ial Superiority in Severe ET Effect Using Hie rarchical Regr essions ............................................. 107 Overview ........................................................................................................ 107 Analyses ......................................................................................................... 108 MRT (180-degrees condition) regressions ............................................... 110 C-G Arena percent time (p robe) regre ssions ........................................... 111 Summary of Results for t he Regression M odels ............................................ 114
7 6 DISCUSSI ON ....................................................................................................... 115 Study Overview and Results Summary ................................................................. 115 Overview ........................................................................................................ 115 Support for Superior Spatial Sk ills in Seve re ET ............................................ 116 Some Unexpected Nu ll Findings .................................................................... 117 Summary ........................................................................................................ 118 Potential Factors Accounting fo r Unexpected Null Findings ................................. 119 Visual Search Task Factors ............................................................................ 119 Judgment of Line Orientat ion Test Fa ctors..................................................... 123 Anti-Tremor Medi cations ................................................................................ 124 Motor Require ments ....................................................................................... 125 Theoretical Assumptions of the Study................................................................... 126 Functional Organization of the Visual Brain .................................................... 126 Behavioral Neuroplasticity Driving t he Spatial Superiori ty Effect? .................. 129 Other Directions for Fu ture Resear ch ................................................................... 131 Significance of the Stud y ...................................................................................... 134 APPENDIX .................................................................................................................. 135 A TREMOR RATING SCALE ................................................................................... 135 B C-G ARENA PARAMETERS AND OBJECT TEXTURES ..................................... 141 Visible Targets Trials (Motor Cont rol Condition): Pa rameters ............................... 141 Visible Targets Trials: Object Text ures ................................................................. 142 Hidden Target Trials (Spatial Memory / Navigation Condition): Parameters ......... 143 Hidden Target Trials: Ob ject Textur es .................................................................. 144 LIST OF REFE RENCES ............................................................................................. 154 BIOGRAPHICAL SK ETCH .......................................................................................... 154
8 LIST OF TABLES Table page 4-1 Rationales for exclusion of participants by group, se x ........................................ 474-2 Demographic and other char acteristics by group ............................................... 495-1 Means (standard deviations [SD]) by group for dementia screening, frontoexecutive, and m ood variabl es ........................................................................... 845-2 Choice Reaction Time ta sk means (SD) by group .............................................. 865-3 Judgment of Line Orient ation means (SD) by group ........................................... 885-4 Mental Rotation Test means (SD) by group ....................................................... 905-5 Visual Search Task means (SD) by gr oup .......................................................... 945-6 Computer-Generated Arena m eans (SD) by group ............................................ 985-7 Facial Recognition Test means (SD) by group ................................................. 1015-8 Object Attention Test means (SD) by group ..................................................... 1045-9 Correlations with Â“Spatial SuperiorityÂ” va riables in essential tremor (ET) ......... 1095-10 Summary of hierarchical regression analyses for variables predicting spatial measures in ET ................................................................................................ 113
9 LIST OF FIGURES Figure page 4-1 Overview of the te sting session procedur e ......................................................... 434-2 Upper extremity Tremor Rating Scale (TRS) score distributions for the mild essential tremor (ET) and severe ET groups ...................................................... 494-3 Input modalities for t he computeriz ed tasks. ...................................................... 554-4 Choice Reaction Time (CRT) task st imuli. .......................................................... 584-5 Mental Rotation Task (MR T) example st imuli. .................................................... 614-6 MRT example stimuli, grouped by increasing degree of Â“biomechanical constrai nt" ........................................................................................................... 644-7 Visual Search Task ( VST) example stimu li ......................................................... 674-8 Computer-Generated Arena (C-G Ar ena) example views as seen by participant s ......................................................................................................... 704-9 C-G Arena example path plots depi cting spatial navigation and memory acquisiti on .......................................................................................................... 744-10 Object Attention Task (OAT) example stimu li ..................................................... 765-1 CRT performance by group, respons e hand ...................................................... 865-2 Judgment of Line Orientation task raw scores and ageand gender-corrected scores ................................................................................................................. 885-3 MRT performance by group (control, mild ET, severe ET), performance measure (RTs and error rates), and condition (angle of rotation) ....................... 905-4 VST reaction times and error ra tes, by group and condit ion ............................... 945-5 Motor control on C-G Arena (Â“v isible targets trialsÂ”) ........................................... 985-6 C-G Arena Â“hidden target tria lsÂ” performance by gro up ...................................... 995-7 Facial Recognition Test raw score s and ageand education-corrected scores 1015-8 OAT reaction times and e rror rates by group ................................................... 104
10 Abstract of Dissertation Pr esented to the Graduate School of the University of Florida in Partial Fulf illment of the Requirements for t he Degree of Doctor of Philosophy ESSENTIAL TREMOR: PARADOXICAL VISUOSPATIAL COGNITION By Utaka S. Springer August 2010 Chair: Dawn Bowers Major: Psychology Essential tremor (ET) is the most co mmon tremor disorder in humans and is associated with older age. In this condition, tremor manifests from pathological processes involving the cerebe llum and primarily occurs durin g intentional movement of the hands and arms. When severe, it can le ad to dysfunction, even impairment, in activities of daily living. ET is also associated with non-motor symptoms, and a handful of studies addressing cognitive functioning in ET have emerged within the past decade. While these studies have consistently descr ibed mild deficits in fronto-executive functions (i.e., attention, verbal fluency, se t-shifting), visuospatial functioning has been described much more variably. The present study is the fi rst to focus on visuospatial functioning in ET with a hypothesis-driven, experimental approach. Based on a preliminary neuropsychological study and other structural neuroimaging data, it was hy pothesized that visuospatial functioning in ET is related to a counterintu itive clinical phenomenon, i.e., that greater tremor severity in ET is predictive of better spatial cognition. Although progressive neurological conditions are generally associated with cognitive dec line, it is feasible that patients with more severe in tention tremor of the hands tend to develop better spatial
11 abilities via various compensatory mechani sms, e.g., greater attention needed for continuously judging and monitoring spatial as pects of the environment (particularly distances between the hands and nearby obj ects for use or avoidance). This hypothesis was tested in the pres ent study with cognitive / behavioral measures. Individuals with ET were recrui ted and categorized into a severe or mild tremor group, then tested co mprehensively along with contro ls on various clinical and experimental measures of visual cognition (spatial tests and object/form tests). Results showed that the severe ET group indeed outperformed the mild ET group on two measures of spatial ability, a mental rota tion measure and a spatial memory/navigation measure. The two groups were compar able in demographic and mood variables, and the mild ET group demonstrated better fronto-executive scores than the severe group. Neither patient group perfo rmed better than normal controls on any measure. Regression analyses showed that tremor seve rity in ET positively predicted better mental rotation ability and spatial me mory, above and beyond other demographic, cognitive, and mood predictor variables. These findings support the notion that compensation for tremor in ET is associated with improvement spatia l skills, particularly on tasks with action-based, dynamic components.
12 CHAPTER 1 OVERVIEW AND AIMS Essential tremor (ET) is the most common age-associated neurological tremor disorder in the United States. In this c ondition, tremor (predominantly in the hands and arms) manifests during purposeful movement s and can impair daily activities when severe. Disturbances in cognitive functioni ng also have been described in the emerging literature, particularly mild deficits in fronto-executive functioning (e.g., set shifting, speeded fluency). Visuospatial f unctioning, the focus of the present study, has been described much more variably. Some studies have demonstr ated relatively intact performance on visuospatial tests, while others have described m ild impairments. Other evidence suggests that visuospatial functioning in ET may be unusually high. The basis for these divergent visuospatial findings is unclear. One possibility is that of methodological variance across studi es. Usually, only one or two tests have been used to study visuospatial functioning in small, ill-characterized samples, often with no specific hypotheses or predictions. Mo reover, a wide variety of tests have been used to study visuospatial functioning, so t he divergent findings in this literature may be a function of the type of test used. Some studies lack appropriate attention to screening methods for general cognitive impairm ent, and many others have used visuoconstructional tests that requi re efficient fine-motor manipul ations of objects (e.g., a pencil or small blocks), potent ially problematic for patients with a tremor disorder. For these reasons, prior studies likely have underes timated visuospatial functioning in ET. Beyond poor methodological control, t he divergent characterizations of visuospatial functioning in ET might be related to some systematic, clinical phenomenon, such as markers of disease severity (e.g., severity of tremor). While one
13 might expect that greater disease severity would be predictive of poorer visuospatial ability, a preliminary study found surprisingly hi gh spatial abilities among the majority of a group of relatively older and severely impa ired pre-surgical ET patients (19 of 22) (Springer, Chang, Graf-Radford, Jacobson, Crucian, Okun et al., 2006). A separate study of another ET sample and controls yielded structural neuroimaging data that complement these behavioral findings. Specif ically, areas of the brain used in spatial processing were found to be relatively dens e in ET patients wit h more severe and longstanding intention tremor of the arms; however, ET pat ients with little or no upper extremity action tremor showed typical age-re lated decline in these visuospatial areas of the cortex (Daniels, Peller, Wo lff, Alfke, Witt, Ga ser et al., 2006). Taken together, these findings are suggestive of a paradoxical phenom enon: more severe disease (i.e., upper extremity intention tremor) may be related to better visuospatial functioning in ET. The behavioral data and structural neur oimaging data of these previous studies also suggest that this relationship (if legitimate) may hold only for judgments about spatial relationships, but not other aspects of visual cognition. This proposed relationship is counterin tuitive and necessitates the suggestion of an explanatory mechanism. Feasibly, patient s with more severe u pper extremity tremor need to recruit greater compensatory skills to maintain similar functional levels. These skills likely involve visuospatial functions (e .g., spatial judgments, vi suomotor tracking), because patients with intention tremor are affected primarily while reaching for, manipulating, or avoiding near by objects. Over time, pat ients likely undergo Â“practiceÂ” of these skills, leading to m easurable improvements on cogni tive tasks and neuroplastic changes in spatial areas of the brain.
14 The overall aims of this study were twofol d. The first aim was designed to test the hypothesis that relatively severe upper a rm tremor in ET is associated with better visuospatial skills relative to ET patients with milder arm tremor. This hypothesized relationship was constrained to spatial ta sks only, and not those requiring judgments regarding feature recognition/discrimination (paralleling the theor etical distinction between spatial and object/form visual st reams of information processing by Ungerleider and Mishkin, 1982). The second aim of this study was designed to address and rule out other variables contributing to visuospatial functioning, should a counterintuitively positive relationship bet ween disease severity and visuospatial functioning be demonstrated in the preliminary analyses. More specifically, the contribution of disease severity to spatia l abilities in ET was tested, above and beyond other sociodemographic, cognitive, and mood predi ctor variables. To meet these aims, various clinical and experimental tasks r equiring different types of spatial and object/form processing were administered to individuals with varying degrees of tremor severity, as well as to normal controls. Neither of the ET gr oups was expected to outperform the controls on ei ther the spatial tasks or the object/form tasks. The following chapter provides a detailed theoretical backdrop for this study, including sufficient background for ET as a disorder, a discussion of behavioral and structural plasticity in neurological diseas e and older age, and a discussion of subtypes of visuospatial processing. The specif ic aims and hypotheses are detailed in the subsequent chapter (Chapter 3).
15 CHAPTER 2 BACKGROUND AND SIGNIFICANCE Essential Tremor (ET) The central topic of this dissertation is visuospatial functioning in essential tremor (ET). This chapter provides background informa tion relevant to this topic and highlights the importance of this line of research. Pr ior to a discussion of visuospatial functioning in ET, however, it is necessary to provide an overview of the disorder itself. Therefore, this initial section of this chapter briefl y covers the presentation, impact, neuropathology, and treatment of ET. ET is recognized as the most common neur ological tremor disorder in humans (Rautakorpi, Takala, Marttila, Sievers, & Rinne, 1982; Benito-Leon & Louis, 2006). It has been estimated as 20 times more common than ParkinsonÂ’s disease and has a prevalence of about 4.0% in the general population (Dogu, Sevim, Camdeviren, Sasmaz, Bugdayci, Aral, et al., 2003). In people over 65 years of age, the prevalence rate is higher, with estimates ranging bet ween 9% and 14% (Moghal, Rajput, DÂ’Arcy, & Rajput, 1994; Louis, Marder, Cote, W ilder, Tang, Lantigua, et al., 1996). ET is characterized most commonly by b ilateral kinetic/action and postural tremor of the upper extremities, that is, tremor that occurs during voluntary movements or static posing of the arms. While the arms are affected in about 90% of patients, 50% of patients or fewer have tremor in other regions of the bod y, such as the head (neck) and/or voice (Koller, Busenbark, & Miner, 1994) The trunk or legs are affected more rarely and usually after the ons et of arm tremor, suggesting t hat ET initially affects the arms and then spreads to other regions of the body as the disease progresses (Louis, Ford, & Frucht, 2003). Longitudinal studies and anecdotal reports indicate that the
16 progression of ET tends to be relatively gr adual, although the course of the disease can vary considerably across individuals (Elble 2000; Putzke, Whaley, Baba, Wszolek, & Uitti, 2006). Among those with severe tremor, quality of life can be substantially impaired. Levels of tremor severity in ET have not been formally defined for widespread clinical use, but relatively severe tremor of the arms, for example, can interfere with basic activities of daily living (ADLs), e.g., hand ling a spoon or glass, buttoning a shirt, or using soap or a toothbrush (Lewis, 2002). Un fortunately, disability does not affect a minority of individuals with ET. Almost three out of four ET patients report some type of disability due to tremor, and many are forced to quit their profession due to disabling shaking (Louis, Barnes, Albert, Cote, Schneier, Pullman et al., 2001). Converging data from imaging, lesions and neuropathology studies suggest that the characteristic action tremor of ET is the result of patho logical processes localized in the cerebellum or its inflow (ponto-cerebellar) or outflow (cerebello-thalamic-cortical) pathways (Louis, 2006). Indeed, t he tremor observed in ET is very similar to that seen after damage to the cerebellu m, and cerebellar stroke can completely abolish tremor symptoms altogether (Dupuis Delwaide, Boucquey, & G onsette, 1989). Positron emission tomography was used to show that at rest, cerebellar areas are hyperactive by 30-40% in ET patients relative to healthy co ntrols (Wills, Jenkins, Thompson, Findley, & Brooks, 1994). A relatively large, case-cont rolled, post-mortem st udy found widespread Purkinje cell loss and a seven-fold increas e in cerebellar Â“torpedoesÂ” (clusters of misaccumulated, disorganized neurofilaments and ot her organelles) in the majority of ET patients (Louis, Faust, Vonsattel, Honig, Rajput, Robinson, et al., 2007).
17 First-line medical treatment for easing or negating tremor typically begins with pharmacological agents. The more popularly prescribed and empirically supported agents include beta-adrenergic antagonists (e.g., propranolol) and anticonvulsants (e.g., primidone), although other agents are used as well, including benzodiazepines, botulinum toxin, and atypical antipsychotics (Pahwa, Lyons, & Pahwa, 2005; Zesiewicz, Elble, Louis, Hauser, Sullivan, Dewey, et al., 2005). The majority of patients do not experience major side effects, so medications can substant ially improve the quality of life by reducing tremor and associ ated impairment ( Louis, 2006). Neurosurgical treatment is available fo r medication-refractory cases. These include thalamotomy or deep br ain stimulation surgery of the ventral intermediate nucleus of the thalamus (Lyons & Pahwa, 2004; Okun, Rodriguez, Mikos, Miller, Kellison, Kirsch-Darrow, et al., 2007; Pahw a, Lyons, & Koller, 2000). Both of these procedures are designed to e liminate or lessen tremor by interrupting hyperactive pathways from the cerebellum to fronto-motor pathways via modification of thalamic functioning. Due to the risk of potentially i rreversible adverse effe cts inherent in brain surgery, these procedures ar e limited to the disabling case s that do not show sufficient response to medications. Cognitive Functioning in ET Aside from the primary moto r symptoms of ET, mild cognitive changes also have been reported. Neuropsychological studies of ET have emerged in the past decade and have consistently described mild deficits in frontally mediated executive functions such as cognitive inhibition, verbal fluency, and set-shifting (e.g., Benito-Leon, Louis, & Bermejo-Pareja, 2006a; Duane & Vermilion, 2002; Gasparini, Bonifati, Fabrizio, Fabbrini, Brusa, Lenzi et al., 2001; Higgi nson, Wheelock, Levine, King, Pappas, &
18 Sigvardt, 2008; Lacritz, Dewey, Giller, & Cullum, 2002; Lombardi, Wollston, Roberts, & Gross, 2001; Sahin, Terzi, Ucak, Yapici, Basoglu, & Onar, 2006; Springer, Chang, GrafRadford, Jacobson, Crucian, Okun, et al., 2006; Troster, Woods, Fields, Lyons, Pahwa, Higginson et al., 2002). These deficits hav e been observed in severe, presurgical patients and community dwelling populations and have been attributed to dysregulation of frontal circuits via hyperactive cerebel lar-thalamic outflow pathways. Side effects from high levels of tremor-reducing medica tions (e.g., primidone, propanolol) also have been potentially implicated in these iss ues; however, relatively young and newly diagnosed Â“treatment-naveÂ” ET patients also have been shown to manifest subtle fronto-executive problems, ra ising doubts as to whether me dications are the primary basis for cognitive difficulties in ET (Sahin et al., 2006). A Preliminary Study: Surpri sing Visuospatial Findings At the University of Florida, Sprin ger and colleagues retrospectively examined data from an older and relatively severe group of ET patients (N=24), who was evaluated prior to undergoing DBS surgery thr ough the UF Movement Disorders Center (Springer et al., 2006). The underlyi ng rationale was to determine how the neuropsychological profiles of these severe, ol der ET patients would differ from those of individuals with ParkinsonÂ’s disease. As a first step, the cognitive profile of ET alone was analyzed in an attempt to replicate findings from the literature. It was hypothesized that frontal-executive impairments would also be observed in this severe population, even after eliminating patient s from our sample for probable dementia, severe mood disturbances, brain trauma, and little formal educ ation (< 9 years). Neuropsychological measures were used to test several c ognitive domains, including memory, language, attention, fronto-executive functioning, and visuospatial f unctioning. Analyses were
19 designed to examine the propor tion of patients experiencing cognitive deficits, with the final sample of pre-surgical patient s averaging 70 years of age and 14 years of education. Consistent with other reports in the literature, the resu lts of these analyses showed that a disproportionate number of ET pat ients performed poorly on tasks associated with frontal-executive functioning (i.e., measures of inhibition and speeded verbal fluency). Results also indicated, however that an abnormally large proportion of the sample performed well on the Judgment of Li ne Orientation (JOLO) task, a clinical visual-spatial task requiring fine discriminati ons of linesÂ’ angular or ientations (Benton, Hamsher, Varney, & Spreen, 1983). Specif ically, 19 of the 22 ET patients who completed this test scored in the 51st to 100th percentile range relative to ageand gender-corrected norms. As a group, the ET patients scored about half a standard deviation above the 50th percentile. This elevated performance occurred in the context of average memory performance (stories, word list) and another visuoperceptual test of object discrimination (unfamiliar face matching and recognition), in addition to the mild fronto-executive functioning deficits. The higher-than-expected visuospatial func tioning in this severe ET population was a surprising finding. Certainly, previous studies examining visuospatial cognition in ET have shown variability, from descriptions of deficits (e.g., Duane & Vermilion, 2006) to Â“intactÂ” abilities (e.g., Lombardi et al., 2001). Howeve r, neurological conditions are not typically associated with relatively be tter or improved cognitive abilities. In ParkinsonÂ’s disease, for example, there is evidence to suggest that impairment in visuospatial processing is among one of the most frequently reported cognitive
20 complaints (Growdon & Corkin, 1987). Moreov er, the high visuospatial performance in this sample was found the context of mild deficits on fronto-executive and other tests. This suggests that the sample showed exec utive functioning characteristics comparable to other study samples and was not simp ly gifted across cognitive domains. Methodological Concerns The basis for divergent visuospatial findings in the ET literature is unclear and is a central question in this dissertation. T here are several possible explanations. One relates to methodological varianc e. Across cognitive studies in ET, no single test has served as a Â“common denominatorÂ” to assess vi suospatial functioning. Because many visuospatial tests are Â“multifactor ialÂ” (in that they are sensit ive to variations in factors beyond primary visual processing, especially in tremor disorders), divergent findings across studies might be a product of the type of test or tests used (Waterfall & Crowe, 1986). For example, many te sts that are commonly used to measure visuospatial functioning in ET require speeded and comple x fine-motor responses, such as copying a detailed line drawing or manipulating small, colored blocks to replicate a target pattern within a certain time limit, e.g., the Rey-Os terrieth Complex Figure Test Â– Copy Trial (Duane et al., 2002; Kim et al., 2009; Sahin et al., 2006), or the Block Design subtest from the Wechsler Adult Intelligence Scale, Th ird Edition (Higginson et al., 2008; Lacritz et al., 2002; Sahin et al., 2006; Wechsler, 1997a) Individuals with relatively severe tremor may have a more difficult time on these types of tests, leading to an overall bias toward describing impairments in these studi es. Test-specific differences might also occur on non-motor based tasks, simply as a re sult of differences in test difficulty. Additionally, not all studies in ET have used adequate screening methods for conditions associated with general cognitive impairment, such as dementia, severe
21 psychiatric impairment, or a history of se vere brain trauma. This too may have contributed to an overall bias toward descrip tions of visuospatial impairment (as well as impairment in other cognitive domains). Severa l studies do not report the details of their screening methods, or they faile d to mention screening for these conditions altogether (e.g., dementia, as in Lombardi et al., 2001, or Duane & Vermilion, 2001). Because ET has been associated with higher rates of dement ia and psychiatric disturbances in some studies (Benito-Leon, Louis, & Bermejo-Parej a, 2006a; Benito-Leon, Louis, & BermejoPareja, 2006b; Schneier, Barnes, Albert, & Loui s, 2001), screening for these conditions is especially important for clarifyi ng visuospatial functioning in ET. Â“SpatialÂ” vs. Â“Object/FormÂ” Processing: Subtypes of Visuospatial Cognition Another consideration for the divergent vi suospatial findings in the ET literature relates to theoretical subtypes of visual co gnition. It is true that current functional models of the visual brain remain open to debate (Milner & Goodale, 2009), and Carroll (1993) suggested after his comprehensive review of factor-analytic visuospatial studies that the actual number of visuospatial Â“abi litiesÂ” may equal the number of different visuospatial tests created. Regardless, a f unctional model of visual processing may be useful in understanding visuospat ial findings in ET studies. In one highly influential model, Ungerleid er and Mishkin (1982) proposed that the raw visual information that reaches the primar y visual cortex from the retinal fields bifurcates and is processed further along two separate visual streams, each with separate functions. They posit ed that the anatomically more ventral processing stream (colloquially dubbed the Â“whatÂ” pathway) proj ecting to the inferior temporal lobes processes relatively static form features for the purpose of object identification (e.g., facial recognition, or form discriminati on). The other processing stream, they
22 suggested, runs dorsally to the parietal lobes (i.e., the Â“whereÂ” pathway) and was theorized to mediate relatively spatial aspects of visual cognition, allowing the perceiver to localize elements of the environment. Th e forward projections of the two visual streams of processing are relatively segregated before conver ging in widespread cortical areas, where information is integrat ed and utilized in higher-order cognition. Areas of convergence inclu de the superior temporal ar eas and various frontal and prefrontal locations (Gol dman-Rakic, 1987; Ungerleider Courtney, & Haxby, 1998). In essence, the model of Ungerleider and Mishkin (1982) suggested that the Â“purposeÂ” of vision is twofold: to recogni ze and localize elements of the surrounding world. Indeed, this intuitive notion has been supported by several lesion, functional imaging, and electrophysiology studies, wh ich have demonstrated evidence consistent with the idea that the percept ion of spatial relationships and the perception of objectsÂ’ forms indeed occurs via distinct neuroanatomical substrates; i.e., an occipito-parietal stream of visual information processing, and an occipito-temporal route, respectively (Ungerleider & Haxby, 1994). Factor analytic studies also have highlighted a primary distinction between spatial cognition and object/form perc eption based on performances in brain-injured and healthy populations acro ss a wide variety of visuospatial measures (for review, see Carroll, 1993). While not free from controversy itself, the theoretical distincti on between subtypes of visual processing might help account for the disparate descriptions of visuospatial functioning in ET. In Springer and coll eaguesÂ’ (2006) study, for example, two visuospatial tests were used, one traditionally considered to measure spatial/dorsal stream functioning (i.e., Bent onÂ’s Judgment of Line Orient ation), and one traditionally
23 considered to measure object form/ventral stream functioning (i.e ., BentonÂ’s Facial Recognition Test). The aut hors found that their severe, pre-surgical ET sample performed as expected on the fa cial discrimination task but better than expected on the spatial task. The dissociation may be rela ted to differences in functioning between Ungerleider and MishkinÂ’s (1982) two visual str eams of processing. It is worthy to note that the pattern observed by Springer et al (2006) was found in a well-screened, severe ET sample, and with the aid of clinical te sts lacking a motor response requirement. A Hypothesis for Superior Spatial Cognition in ET The basis of disproporti onately high spatial functi oning (but not object/form perception) found in severe ET is unclear and seems paradoxical (Springer et al., 2006). A potential explanation might be relat ed to findings from a recent structural neuroimaging study. Daniels and colleagues (2006) used vo xel-based morphometry to test the hypothesis that severi ty of kinetic/action tremor in ET may be associated with cerebellar degeneration. Results failed to confirm this hypothesis, as cerebellar volumes were not found to vary by diseas e severity when a Â“severe ETÂ” sample was compared with a relatively Â“mild ETÂ” sample and normal controls (disease severity was defined by intention tremor sco res for the arms). The author s did incidentally find that despite the more severe ET group being slig htly older than the mild ET group, the severe group had relatively larger cortical volumes in inferior parietal cortex, the posterior region of the superior temporal gyrus (pSTG), and the parahippocampal gyrus, with right-sided volumes being greater t han on the left, when compared to age-matched controls. These areas of the brain have been shown to play vital roles in spatial processing. The parietal lobes have been implicated as be ing important for spatial attention and
24 manipulation of mental images, as in m ental rotation tasks (Jordan, Heinze, Lutz, Kanowski, & Jancke, 2001; Zacks, 2008). On t hese types of tasks, individuals must mentally spin or rotate objects to make accurate perceptual judgments. Reversible lesion studies (i.e., using transcranial magnetic stimulation and direct cortical stimulation) have found t hat the right pSTG is critical fo r some types of visual search tasks (i.e., serial search for targets perceptually similar to other objects in the array) (Ellison, Schindler, Pattison, & Milner, 2004; Gharabaghi, Berger, Ta tagiba, & Karnath, 2006). Lesion and neuroimaging studies hav e consistently found evidence that the parahippocampal gyrus, especially on the ri ght, is important for spatial memory and navigation (Aguirre, Detre, Alsop, & DÂ’Es posito, 1996; Barrash, Damasio, Adolphs, & Tranel, 2000; Epstein, DeYoe, Press, Rosen, & Kanwisher, 2001). In effect, the areas of cortex that Daniels and co lleagues (2006) found to be relatively denser in Â“severe ETÂ” patients relative to ET patient s with milder arm tremor are re latively dorsally located in the brain and have been shown to be important substrates for s patial cognition, including mental rotation, visual search, and s patial memory/navigation. This raises the possibility that patients with severe ET mi ght perform better on spatial tasks measuring these abilities. From a neuroanatomical basis, the findings of Daniel s et al. (2006) do not suggest that disease severity is predi ctive of object/form processing abilities mediated by more ventral occipito-temporal ar eas, e.g., face processing (see review by Young, De Haan, & Bauer, 2008). As a side note, because mental rotations, visual search, and spatial memory/navigation will be assessed experimentally in this study, more detailed background on these constructs will be provided in a later section of this chapter.
25 The findings from Daniels and colleagues (2 006) also suggest that ET patients with milder arm tremor have less cortical density in spatial processing substrates relative to patients with more severe arm tremor. Their st ructural data suggest that patients with mild ET woul d demonstrate measurably wo rse performance on spatial tasks relative to those with more severe ET In line with this view, Sahin and colleagues (2006) evaluated a group of newly diagnosed ET patients with milder tremor who had not yet begun medications. Relative to controls, these young ET patients had consistently poor performance across three clin ical tasks of visual cognition (i.e., face discrimination, judgment of the orientations of lines, construction of block designs). Altogether, a positive relationship might ex ist between intention tremor severity and spatial cognition in ET, as inferr ed from these behavior al and structural neuroimaging findings. The explanatory mec hanism for this relationship remains open to conjecture. It is assumed that t he effective coordinati on of tremulous limb movements in daily life requires effort proporti onal to the severity of the tremor itself. The positions of the hands relative to objec ts intended to be used (or avoided) must be tracked continuously during use, and with mo re severe tremor, more attentional resources would need to be diverted to co mpensatory cognitive functions, such as spatial computations. Over time, because individuals with se vere kinetic tremor of the arms exercise these abilities daily as they co mpensate for their tremor, it is feasible that their spatial abilities improve measurably, as demonstrated by superior performance on visuospatial tests and greater density in spatial cortex via neuroplastic changes. Neuroplasticity and Visuospatial Cognition Neuroplasticity is the brainÂ’s ability to undergo structural alterations and functional adaptations in response to internal c hanges (e.g., CNS injury, disease) or
26 environmental changes (e.g., experience, tr aining, or living with tremor). This mechanism has been proposed to account for such findings as the positive correlation between posterior hippocampal volumes and driving experience / spatial navigation ability in London taxicab drivers (Maguire Gadian, Johnsrude, Good, Ashburner, Frackowiak, et al., 2000), as well as the revers ible effect of only a few monthsÂ’ juggling training on the size of motion-processing ar eas in visual cortex (Draganski, Gaswer, Busch, Schuierer, Bogdahn, & May, 2004). N europlasticity may involve spine density changes or neurogenesis (Grutzendler, Kasthur i, & Gan, 2002; Trachtenberg, Chen, Knolt, Feng, Sanes, Welker, et al., 2002; Kempermann, Kuhn, & Gage, 1997), and in the context of slow-progre ssing CNS damage (e.g., low-gr ade gliomas), it is usually associated with complete or near-complete reco very of cognitive functions (Desmurget, Bonnetblanc, & Duffau, 2007). Although the progression of ET symptoms varies individually, ET is generally associated wit h a slow rate of advancement with marked changes often on the order of decades. Sp ringer and colleaguesÂ’ (2006) observation of better-than-average visuospatial performanc e in ET is compatible with these observations. Moreover, there is evidence that vis uospatial functions are among the most functionally plastic and amenable to rehab ilitation after brain injury. Cicerone and colleaguesÂ’ review of 87 cognitive rehabilitati on studies found that the most rigorous randomized controlled studies have demonstrated positive and lasting benefits for visual inattention (i.e., hemispatial neg lect) and cortical blindness interventions, whereas such evidence is lacking for non-strategy based Â“executiveÂ” skills (Cicerone, Dahlberg, Malec, Langenbahn, Felicetti, Kneipp, et al., 2005). Benefits for training aspects of
27 visual attention have been demon strated not only for brain injured populations, but for the healthy young and the healthy elderly as well (e.g., Ball, Beard, Roenker, Miller & Griggs, 1988; Ball, Berch, Helmers, Jobe, Leveck, Marsiske et al., 2002; Edwards, Wadley, Myers, Roenker, Cissell, & Ball, 2002; Kramer, Bherer, Colcombe, Dong, & Greenough, 2004). Other visuospatial skills appear to benefit greatly from training, such as mental imagery or rotati on ability (Lohman & Nichols, 1990). Moreover, it has been established that training effe cts can generalize or Â“transferÂ” to other skills, such as video game training effects on selective visual attention or speed of training effects on instrumental ADLs (Edwards et al. 2002; Green & Bavelier, 2003). The domains of memory and fronto-executive functioning have not shown such amenability to functional plasticity, perhaps accounting for the discrepancy between superior visuospatial functions but mildly impaired fronto-executive functions in severe ET patients (Springer et al., 2006). Regardless, a synthesis of t hese ideas suggests that int entional tremor in severe ET may be associated with superior spatial abi lities. Functional neuroplasticity is commonly observed in the visuospatial domain of cognition for healthy and braindamaged individuals, it has been documented with substantia l and lasting change in patients well above 65 years of age, and robust maintenance (or improvements) in performance can be observed when damage is subt le in progression, as is observed in ET. Tremor Severity vs. Tremor Duration: Potential Visuospatial Predictors The principles derived from the neuroplasticity literature raise the question as to whether tremor severity (i.e., amplitude) or tremor duration might better predict visuospatial performance in ET. Because of individual variability and the overall slow
28 progression of ET severity, these two constr ucts are not highly correlated in ET (Elble, 2000; Putzke, Whaley, Baba, Wszolek, & Uitti, 2006). It was suggested above that neuroplastic changes may occur in visuospatial co rtex as a result of a chronic increase or diversion of attentional resources to vis ual spatial elements of the environment. The question arises, however, as to whether it is time or the Â“doseÂ” (i.e., amplitude of the tremor to which patients must adjust) that is associated with the mo st change. Both are likely necessary, but it is argued here that the dose, or severity of the tremor, is the more important factor. It is likely that t he amplitude of the tremor must fall above some threshold level in order to elicit more effo rtful engagement in visual attentional resources and practice of spatial computational function s. Conversely, because very mild tremor often presents negligible difficulty to the indivi dual, it likely requires li ttle if any functional adjustment. As neuroplastic changes require some internal or environmental change, it is unlikely that very mild tremor would be associated with improvements in behavior or cognition. Furthermore, it has been established that neuroplastic changes in visuospatial cortex can occur in healthy, younger individu als after only a few months of training (Draganski et al., 2004). This timeline likely is much faster: rats t hat were trained on a skilled reaching task showed increases in synaptic strength, sy napse number, and map reorganization after only several days of training, and this occurred after making significant behavioral gains, as reviewed by Kleim & Jones (2008). Granted, neuroplastic changes are slower with increasi ng age, but at any given amplitude of tremor to which the individual must adjus t, the duration of ti me after a few years (conservatively) would likely not have as much association with neuroplastic changes as
29 the amplitude of the tremor itself (Kleim & Jones, 2008) This may be the case especially because the rate of progression in ET tends to be slow; significant increases in tremor severity do not occur on the or der of months but ra ther, years or even decades, whereas the rate of neuroplastic ch anges tend to occur on a much faster scale. Finally, tremor severity often is a more reliable measure than di sease duration. It can be assessed in an entire research sample by a single clinician, whereas disease duration depends on the reliability of each patient Â’s self-reported histor y. Many patients report being unaware of tr emor symptoms until a doctor noti fied them of their presence. Other patients report first noticing tremor symptoms but are unable to recall precisely when in their lives this occurred. Thus, es timates often are given with a 5 or 10 year margin of error. Taken together (and the degree to which they can be measured reliably), tremor severity ra ther than duration of tremor symptoms is likely associated more with changes in spatial cognition and it s associated neurological structures (if such associations exist). Spatial Cognition Subtypes: Foci of the Present Study As was reviewed above, the structural neur oimaging data of Dani els et al. (2006) indicate that ET patients with more severe upper extremity tremor (relative to those with mild or no tremor) have greater cortical volumes in parieta l cortex, posterior superior temporal gyrus, and posterior parahippocampal gyrus (with volumes being larger on the right side). These three areas have been asso ciated with different types of spatial abilities, including mental rotation, visual sear ch (hard-feature serial search), and spatial memory and navigation, respecti vely. For this reason, indi viduals with relatively severe ET vs. milder ET might have Â“neuroanatomic al advantagesÂ” on these types of tasks and
30 perform measurably bette r due to compensational differenc es. This chapter concludes with relevant background on these subtypes of spatial cognition and their measurement in order to provide sufficient background for the present study. Mental Rotation Â“Mental rotationÂ” refers to t he ability to spin or rotate mental representations of objects. In the classic m ental rotation experiment, S hepard & Metzler (1971) asked participants to compare drawings of threedimensional geometric st imuli to images of those stimuli rotated about an axis. They were asked to judge whether the second (rotated) stimulus was the same as the first (fixed) stimulus, or w hether it represented a mirror image of that stimulus. Results of this study and many others that followed indicated that reaction time for this task incr eased in linear propor tion to the angle of disparity between the rotated object and the fixed object. Moreover, the accuracy of judgments on these tasks decreased as a functi on of the angle of disparity between the two objects. These findings have been repr oduced extensively over the past several decades, and they suggest that mental im ages in the brain are maintained and manipulated as Â“topographi c wholes.Â” Functional imaging and lesion studies have provided a wealth of evidence suggesting that the parietal lobe, especially in the right hemisphere, plays an important role in the processes of mental imagery and ro tation. For example, parietal activity has been shown to vary positively and parametrica lly by the difficulty (or distance of) the mental rotation performed (Zacks, 2008). Bec ause Daniels et al. (2006) found parietal regions (especially on the right), to be more co rtically dense in severe ET in comparison to milder ET, there may be a neuroanatomical basis for the proposition that severe ET patients might perform better on mental rotation tasks.
31 Relevant to the present study, mental rotation with hand-based stimuli may be a skill in which relatively severe ET patients ar e especially proficient. In day-to-day life, oneÂ’s hands are used to interact skillfully wit h objects in the environment, most often with visual guidance. ET, a ffecting primarily the voluntar y movement of the hands and arms, impairs the ability to interact with objec ts. It is likely that as a compensatory mechanism of adjustment, severe tremor in ET requires effortful attention and use of spatial skills relating to visuomotor pl anning and manipulation of the hands. A handbased mental rotations task seems to be a logical candidate for demonstrating these changes in ET, if they exist. The use of hands as stimuli in mental rotation studies does change the operating characteristics of these tasks to some degree, however. Traditionally, inanimate geometric objects such as twoand three-di mension block figures are used as stimuli (e.g., Shepard & Metzler, 1971). However, it is now appreciated that people are faster at mental rotation tasks when body parts or t ools are used as stimuli, particularly when a head, face, or hand is attached to the stim uli (Amorim, Isableu, & Jarraya, 2006). These findings relate to the concept of Â“e mbodimentÂ”, the tendency of an individual to imagine that the particular body part viewed is their own. It has been postulated that when the stimulus to be rotated is a tool, such as a hammer, subjects Â“embodyÂ” the tool; they imagine rotating it with their domi nant hand as opposed to imagining the object rotating by itself in space (Vingerhoets, Lange, Vandemaele, Deblaere, Achten, 2002). Similarly, when hands are used as stimuli in mental rotation tasks, proficiency increases (i.e., accuracy increases, and response time s to generate correct answers decreases) when the hand can be easily imagined as thei r own (embodied). The easier it is to
32 embody a given stimulus, the more effici ently that stimulus can be mentally manipulated. Mental representations of body parts are also more easily formed and manipulated when those body parts are viewed in anatomica lly possible positions. In other words, mental imagery of embodied objects (e.g., hands) is subject to actual biomechanical constraints. The results of studies usi ng mental rotation tasks with hand-based stimuli show that higher error rates and slower re sponse times for correct answers occur in conditions whereby the hand stimuli are viewed at angles that are difficult to produce in reality (e.g., viewing the back of oneÂ’ s own hand with the fingers pointing down) (Amorim et al., 2006). Speed of rotation is faster and accuracy is higher when hand stimuli match more natural positions (e.g., a view of the back of the hand with the fingers facing away from the body). For t hese reasons, images of rotated hands are harder to mentally rotate when they are pres ented at certain angles of rotation relative to the reference point. The specific task used in the present study is described in the Methods section (Chapter 4). Visual Search Visual search is an intrinsic and vital part of everyday experience for non-visually impaired people. Finding particular objects by sight might be an especially important skill for individuals with severe ET. For example, an increased awareness of objectsÂ’ locations nearby might be necessary for avoidi ng accidents or injury or damage to the objects themselves resulting from motor cont rol difficulties. Visual search occurs relatively automatically or with more sustai ned effort. For exampl e, if an object has a very distinctive feature, it might visually Â“pop outÂ” of the scene to capture oneÂ’s immediate attention, as in spotting a red card inal in the snow among several sparrows.
33 If an object is located in the midst of objects that share more visual features, then it might be found only after relative ly effortful, serial search ing (e.g., finding oneÂ’s house key among other keys on the ring). Visual search tasks have been developed to study these and other attention phenomena both clinically (as in the assessm ent of hemispatial visual neglect), and experimentally (as in understanding the neural bas is of automatic vs. controlled types of visual attention). They are administrat ed in a standardized format using well-defined stimuli, usually arranged on a sheet of paper or a computer screen. Two types of visual search tasks were described by Treisman and Gelade (1980): feature search tasks, whereby the target object is defined by a distinct visual feature (e.g., red, in the cardinal example), and conjunction search, whereby the target object has a unique combination of features shared by other objects in view (e.g., the key, which might be nickel among other nickel and brass keys but have a uniqu e pattern of holes on the grip). The neural correlates of visual search typi cally include the parietal cortex but also depend on the parameters of the visual search task, such as whether attention is overt or covert, or whether the tar get has predefined features or me rely visually distinct from other objects in some way. In the present study, a Â“hardfeature, serial exploratory searchÂ” task was selected because of its reli ance on the superior temporal gyrus (STG), a structure shown to be relatively more dense in severe ET vs. mild ET (Daniels et al., 2006). The role of the right STG on one particular hard-feature search task was demonstrated in two experimental studies. First, Ellison and colleagues (2004) used a stimulation-induced reversible lesion method (trancranial magnetic stimulation) on subjects performing various bisection and vi sual search tasks. A double dissociation
34 was demonstrated: performance on the hard feature-based seri al search task (but not bisection tasks) was impaired when right ST G was stimulated, while impairment was evident on bisection tasks (but not the vis ual search tasks) when the right PPC was stimulated. The right STG was critical only for the hard-feat ure visual search task, but no performance reductions were found on simp le-feature or hard-conjunction visual search tasks. Gharabaghi, Berger, Tatagi ba, and Karnath (2006) supplemented these findings using direct cortical stimulation of the middle-right STG. This group showed that intra-operatively, performance accura cy dropped from near-perfect to chance levels on a modified version of Ellison and colleagues Â’ (2004) hard-feature search task. No decrement in performance was found with stimulation to ot her adjacent areas, however, including the right PPC. These findings sugg est that the right STG is uniquely important for hard-feature serial visual search. Seve re ET patients might be better on this type of task compared with those with milder ET, as the right STG was found to be larger in relatively severe ET. The hard-feature seri al search task used in the present study is described in the Methods section (Chapter 4). Spatial Memory and Navigation The ability to find oneÂ’s way to remembered locations is a vital part of everyday functioning and critical to indepen dent living. When this ability is impaired, as is often the case in early dementia, people may exhi bit wandering and getting lost in familiar places (Klein, Steinberg, Galik, Steele, S heppard, Warren et al., 1999). A spatial navigation / memory task was chosen for the present study because some of the neural correlates for spatial navigation were found to be greater in relatively severe ET vs. milder ET (Daniels et al., 2006) Although it may not be i mmediately clear as to why
35 severe ET would show an advantage in this skill beyond this reason, further discussion of spatial navigation deserves merit to provide context to the present study. Our understanding of spatial memory and na vigation has been informed by a vast literature from animal and human studies. One particularly informative experimental method that comes from the animal literature is the Morris water maze task (Morris, 1981, 1984). In this task, rats are placed into a circ ular pool of opaque water and trained to find a submerged (invisible) platform. When the platform is located in a fixed position, rats find it efficiently, even when released from different locations. Morris (1981; 1984) explained the ratsÂ’ behavior as the consequence of using allocentric strategies, i.e., identifying the position of the platform with respect to its relative position to visual cues outside the pool (the walls of the pool itself had no distinguishing features). The Morris water maze task has been adapted into computerized versions, allowing spatial memory and navigation to be studied in humans in a pragmatic way. One example is the Computer-Generated Arena (Thomas, Hsu, Laurance, Nadel, & Jacobs, 2001). Like the Morris water maze, this particular task also contains a pool of water within a square room, distal cues, and an invisible target, but these elements are displayed virtually in first-person view on a computer monitor. The test subject Â“navigatesÂ” about the pool by using a joyst ick or the keyboard to simulate intended movement, and the first-person view changes according to these movements. This paradigm has been shown to be a practical, reliable, and low-cost method that has successfully replicated the animal literature based on the original Morris water maze.
36 Animal and human data from lesion, st imulation, imaging, and cell recording studies have characterized the neural correlates of spatial navigation. Parietal cortex appears to process visual information in terms of body-centered coordinates. Construction of Â“egocentricÂ” spatial refer ence frames may be used to guide locomotion on spatial navigation tasks from a first-perso n point of view (e.g., Gron, Wunderlich, Spitzer, Tomczak, & Riepe, 2000). More vita l to spatial learning and memory, however, is the hippocampus, which appears to play an integrated role in the processing and storage of allocentric spatial in formation (i.e., locations relative to a spatial reference frame outside the individual) (Ekstrom, Kahana, Caplan, Fields, Isham, Newman et al., 2003; Gron et al., 2000; OÂ’Keefe & Dostrovsky, 1971). Functional im pairment of the hippocampus leads to memory deficits on a variety of maze learning tasks (Morris, Garrud, Rawlins, OÂ’Keefe, 1982), includi ng the Computer-Gener ated Arena (Astur, Taylor, Mamelak, Philpott, & Sutherland, 20 02). The posterior parahippocampal gyrus often called the Â“parahippocampal place area Â” has been shown in many neuroimaging studies to respond preferentially to scenes (images containing information about the layout of a space), rather than other visual ma terial such as objects or faces (Epstein & Kanwisher, 1998; Ekstrom et al., 2003; Gron et al., 2000). A version of the Computer-Generated Arena task was chosen for the present study because some of the neural corre lates for spatial navigation (e.g., parahippocampal and inferior parietal lobe areas, with an apparent rightward asymmetry) were found to be greater in relati vely severe ET vs. milder ET (Daniels et al., 2006). Thus, from a neuroana tomical perspective, it might be predicted that severe ET, relative to milder ET, would dem onstrate better performance on this task.
37 CHAPTER 3 SUMMARY OF THE PROBLEM, SPE CIFIC AIMS, AND HYPOTHESES Summary of the Problem While the emerging cognitive literature in ET has provided fairly consistent evidence for mild frontal-executive functi oning deficits, visual cognition has been characterized much more variably. This in consistency may relate to (1) methodological limitations or variance am ong studies, among them being a failure to distinguish between spatial and object/form subtypes of vis ual processing, and (2) tremor severity as a potential positive predictor of spatial abili ties. First, methodolog ical problems in the cognitive literature in ET relate largely to possible confounds creating a bias toward describing visuospatial deficit s. For example, many visuospatial tasks have a motor component requiring coordinated dexterity of the hands (e.g., WAIS-III Block Design, Rey-Osterrieth Complex Figure). Some i ndividuals with ET might perform poorly on these tasks because severe tremor may prevent effective manipulation of small blocks or a pen. Other studies di d not fully detail or use appr opriate screening methods for general cognitive impairment. Thus, it is unc lear as to whether the ET patients used in these studies were adequately assessed for fact ors that would acc ount non-specifically for visuospatial problems, such as dementia, se vere medical conditions, traumatic injury to the brain, or psychiatric impairment. Aside from methodological flaws, another reason why visual cognition in ET has been difficult to characterize may relate to a failure for studies to dissociate between Â“spatialÂ” and Â“object/formÂ” subtypes of visual processing. The literat ure historically has distinguished between these two modes of visual processing and shown their dissociation in function (i.e., cognitive performance) and neuroanatomical damage. A
38 review of the cognitive and structural neuroimaging ET literature suggests that individuals with ET may be more likely to perform better on tasks requiring spatial perception or judgments, ra ther than on tasks whereby performance is more heavily mediated by visual feature/form processing. Finally, neuropsychological studies of ET have paid little attention to studying factors contributing to spatial cognition in this population. This is possibly because most of the cognitive literature in ET has focu sed on fronto-executive functioning. Recent structural neuroimaging findings by D aniels and colleagues (2006) showed that a subgroup of ET with more severe and slig htly more longstanding arm tremor had relatively denser areas of dorsal-parietal co rtex (right > left) traditionally associated with spatial cognition (e.g., mental rotation, vi sual search, and spatial navigation). The superior spatial performance in ET shown by Springer and colleagues (2006) was also found in a relatively severe (pre-surgical) ET population; this behavioral data appears to complement the structural neur oimaging findings of Daniel s et al. (2006). It is suggested that spatial skills may be Â“exercisedÂ” more in severe ET because of the cognitive compensation required to functionally adjust to this limitation in order to perform daily activities. This may contribut e to improved spatial skills and more robust areas of cortex that mediate these skills. The overall goal of the present study was to further investigate and test the validity of this paradoxical relationship between diseas e severity and spatial cognition in ET. A purely cognitive (behavioral) approach ( and not a neuroimaging approach) was taken in this study as a first step. Various types of visual cognitive ta sks were used under the assumption that the spatial vs. object/form visual cognitio n was an important theoretical
39 distinction, and that subty pes of spatial cognition mi ght have different operating characteristics in this population. The spatia l tasks were also chosen in order to target a potential neuroanatomical advantage in se vere ET, as was suggested by the structural imaging data by Daniels et al. ( 2006). The specific aims of the study are outlined here, with the next chapter describi ng the methods used to carry them out. Specific Aim 1 Overview This study sought to use cognitive test ing as a first step in examining and characterizing visual cognition in ET, usi ng methodologically rigorous methods. Tasks were selected based on theory derived from cognitive neuroscience research, as well as previous findings from cognitive and struct ural neuroimaging research. Both clinical neuropsychological tasks and timed, comput erized experimental tasks were used. Hypotheses In the first specific aim of this study it was hypothesized (paradoxically) that relative to controls, patients with more se vere ET (measured by upper extremity tremor) have better spatial skills than those with le ss severe disease. It was further hypothesized that the severity of ET (high or low) does not have any relationship with visual cognitive skills that are based more on efficient recognition or discrimination of shapes and forms. These tasks were included as control tasks to establish specificity of the superiority effect primarily to the spatial domain only. This study was designed to test these hypotheses while accounting for important methodologic al variables that sometimes have limited the findings of prev ious studies, e.g., appropriate screening methods for generalized cognitive impairment.
40 Predictions It was predicted that a group of severe ET patients (i.e., those with relatively highamplitude tremor) would demons trate superior performance on spatial cognitive tasks relative to an ET group characterized by milder tremor. These spatial tasks involved the judgment of the orientation of lines, mental rotations of ha nds, visual search, and spatial memory/navigation. No group differences were predicted on object/form tasks, i.e., facial recognition or obje ct-based feature di scrimination. Neither ET group was predicted to outperform a group of normal controls. Analyses These predictions were tested via anal ysis of variance (ANOVA) or its nonparametric equivalent (i.e., Krus kal-Wallis), using a betweensubjects factor of group (mild ET, severe ET, controls). Various within-subjects factor s (mixed-models ANOVA) allowed for the testing of c ondition-specific effects by task. Analyses incorporated statistical assumption testing and made appr opriate adjustments when necessary, as detailed further at the end of the me thods section. Specific Aim 2 Overview The second aim of this study, if nece ssary, was designed to address and rule out other potential factors underlying any demonstrat ed spatial superiority effect, in order to more rigorously determine the impact of tr emor severity on spatial functioning and identify the most influential contributor of any spatial superiority in severe ET. That is, the contribution of disease severity to spatia l abilities in the entire ET sample was to be tested, above and beyond other sociodemogr aphic, fronto-executive, and mood
41 predictors. No study spec ifically has examined the relationship between visual cognition and other variables in ET patients. Hypothesis, Prediction, and Analysis In the second aim of this study, it wa s hypothesized that greater upper extremity tremor is predictive of better spatial abilitie s, even after accounting for other correlates of spatial functioning (e.g., age, measures of fronto-executive functi oning, gender, etc.). After combining the mild and severe ET gr oups from the first specific aim into one sample, it was predicted that upper extremity tremor woul d positively predict better spatial functioning (JOLO, mental rota tion of hands, visual search, spatial memory/navigation), above and beyond ot her predictor variables. Spearman correlations were used initially to identify co rrelates of test-specific spatial performance in ET, for variables whereby a spatial super iority effect was f ound (i.e., severe ET outperforming mild ET). These were enter ed in separate stepwise regressions along with the tremor severity variable in order to determine which measures significantly predicted spatial performance. Tremor severity was then tested as a significant predictor of spatial functioning above and be yond any other predictor variables using hierarchical regression, with all other predictors held cons tant in a first model, and tremor severity entered into a second model Final regression models were tested for outliers and meeting statistical assumptions.
42 CHAPTER 4 MATERIALS AND METHODS Design Overview An overview of the testing session procedure is provided in Figure 4-1. As shown, individuals with ET and healthy controls were recruited and screened for various inclusion/exclusion criteria and then underwent a brief neuropsychological battery of attention-working memory, executive (set -shifting, speeded fluency) and reaction time tasks. This battery was followed by clinical and experimental tests of visual spatial and object/form processing. The entire testing procedure took approxim ately 3-4 hours per participant and was conducted at either the Cognitive N euroscience Laboratory at the University of Florida (UF) McKnight Brain Institute or at the participantÂ’s residence (provided that an adequate and quiet test ing environment was available). Approximately equal numbers of participants were tested at home and in the lab, in both the ET and control participant gr oups. Following completion of the study, participants received $20 in cash and reimburs ement for any parking expenses accrued on the UF campus ($3). This study was funded by the American Psychological Foundation (2007 Benton-Meier Neuropsychology Award). Recruitment, exclusion criteria, and testing procedures ar e described in more detail below. Participants Recruitment and Initial Inclusion Criteria A total of 70 individuals (N=36 ET, N=34 Normal controls) between the ages of 50 and 85 years were recruited and completed testing1. To be included in the ET group, 1 Another eight normal controls between the ages of 30 and 50 were tested but were not included in the present study, as no individual with ET who was recruited into the study fell into that age range.
43 Figure 4-1. Overview of t he testing session procedure. Â“SpatialÂ” Tests : Judgment of Line Orientation, Mental Rotation Task, Visual Search Task, Computer-Generated Arena Â“Object/FormÂ” Tests : Facial Recognition Test, Object Attention Test Screening : Dementia screening, medical/ psychiatric/ tremor history, mood/ anxiety measures Â“Fronto-executiveÂ” Tests : WMS-III Digit Span, WMS-III Spatial Span, Stroop Color-Word Test, Letter Fluency (FAS), Category Fluency (Animals), Simp le Reaction Time test (fronto-motor) Recruitment of ET & Control Groups : UF Movement Disorders Center Clinics, ET support groups, retirement communities, support/activity groups for older adults, ongoing studies in the UF Cognitive Neuroscience Lab, community fliers/ads, and word of mouth
44 participants were required to have (a) a clinic al diagnosis of ET by a neurologist, with the primary sign being bilatera l and persistent kinetic trem or of the hands and forearms (Deuschl, Bain, & Brin, 1998), and (b) a stabl e ET treatment regim en for at least one month prior to testing. Participants were assigned to the healthy control group on the basis of a negative history of tremor and fa lling within the age range of the ET patient group. Participants were recruited from several so urces in north-central Florida. For ET participants (total N=36), this included t he UF Movement Disorders Center Clinics (N=21), ET support groups (N=9), Oak Hammock retirement community in Gainesville, Florida (N=3), and via community fliers/ advertisements and word-of-mouth (N=3). Fourteen of the ET participants were candida tes for deep brain stimulation through the UF Fast Track program. Control participants (t otal N=34) were recrui ted from retirement communities in north central Florida (N=12), a local support and activity group for older adults (N=8), ongoing studies in the UF Co gnitive Neuroscience Lab (N=6), community fliers/advertisements and word-of-mouth (N=6), and an ET support group (N=2). Screening and Exclusion Criteria After being recruited into the study, parti cipants provided written informed consent and then were assessed for inclusion/exclusion cr iteria. ET exclusion criteria were as follows: (1) tremor that was unrelated to ET (e.g., sudden-onset, drugs, anxiety), is taskspecific (e.g., writing only), or characteri zed by sudden onset or stepwise progression, (2) report of other abnormal neurological sign s or disease (e.g., dystonia, ParkinsonÂ’s disease), (3) history of moderat e or severe brain injury, e.g ., traumatic injury with a loss of consciousness greater than 30 minutes, neurosurgical interventions involving resection/ablation, history of stroke, (4) current uncontrolled medical illness that could
45 potentially affect cognition (e.g., cardiac or pulmonary disease, hypothyroidism, cancer, HIV), (5) failing screening for basic vision and cognitive impairment / dementia, as outlined below, (6) diagnosis of self-report of si gnificant learning disorder or ADHD, or less than a ninth grade educati on, and/or (7) a history of substance abuse or Â“heavyÂ” alcohol intake2. Exclusion criteria for controls included: (1) evidence of ET (i.e., exhibited or reported a history of tremor) or (2) meeting any of the exclusion criteria for the ET group. Tremor symptoms were assessed usi ng the FahnÂ–TolosaÂ–Marin Tremor Rating Scale (TRS; Fahn, Tolosa &, Marin, 1993). This standardized measure was administered in order to document the presenc e and severity of tremor in the ET group and the absence of tremor in the controls. The TRS is comprised of several subscales that measure three types of tr emor (resting, postural and ki netic) in various regions of the body (e.g., head, arms, legs, etc.), and on various tasks and activities of daily living (ADLs). A scale from 0 (non-ex istent) to 4 (severe) is used for each rating of tremor or impairment level. The range of possible score s on the TRS is 0 to 144. In the present study, the primary measure of tremor severity was postural/ kinetic tremor of the upper extremities (average of left and right scores). Severity scores were also calculated for axial tremor (head, face, t ongue, voice, and trunk), and tr emor-related disability in activities of daily living (speaking, eating, drinking, hygiene, dr essing, writing, and working). The TRS is reproduced in Appendix A. To screen for dementia, all participants re ceived a brief cognitive screening measure called the Mini-Ment al State Exam (MMSE; Folstein, Folstein, & McHugh, 2 Heavy alcohol intake: defined by the National Instit ute on Alcohol Abuse and Alcoholism in the past year (>14 drinks per week for men and >7 drinks per week for women; http://www.niaaa.nih.gov/).
46 1975), a recent memory task in volving immediate and delay ed recall of novel stories (Logical Memory subtests from the Wechsler Memory Scale, Thir d Edition [WMS-III]; Wechsler, 1997b), and a measure of confro ntation naming (the Boston Naming Test [BNT]; Kaplan, Goodglass, & Weintraub, 2001) Individuals who scored below 26 on the MMSE or below the 5th per centile on either the immedi ate or delayed recall portions of the Logical Memory subtest or t he BNT were excluded from the analyses3. A total of six individuals who completed testing were removed after exclusion criteria were applied. They included four males and one female from the ET group, resulting in a final sample size of 31 ET participants (16 males, 15 females). A male control subject also was excluded, resulting in a final sample of 33 control subjects (15 males and 18 females). Thus, the final sa mple used in the statistical analyses consisted of 64 participants (31 ET, 33 controls). The specific reasons for exclusion are shown in Table 4-1. Participant Characteristics The ET sample (N=31) was divided into Â“severe ETÂ” and Â“mild ETÂ” subgroups in order to test the hypothesis t hat individuals with severe vs. mild ET differ relative to controls on spatial visuoperceptual tasks (but not object/form visuope rceptual tasks). At present, there are no standard criteria for distinguishin g between Â“mildÂ” and Â“severeÂ” subgroups in ET. In this study, Â“mildÂ” and Â“severeÂ” ET subgroups were defined on the basis of individual upper extr emity tremor scores (i.e., an average of postural/kinetic upper extremity tremor scores on the TRS for the right and left hands). This measure 3 One participant with ET scored 25 on the MMSE; however, an item analysis revealed that he lost two points due to an illegible written sentence and illegible intersecting pentagons (related to his tremor). His Logical Memory and BNT scores were well within the average range, and therefore, he was not excluded from the study.
47 Table 4-1. Rationales for exclusio n of participants by group, sex. Group Sex Rationale for exclusion Control Male History of learni ng disorder (reading, spelling) Female None excluded ET Male History of learning disorder (reading) Scored in the impaired r ange: WMS-III LM-I & LM-II History of seizures Male Heavy drinking (6 beers/day) Marijuana use (2x/week) Male History of stroke reported Male < 9 years of formal schooling (6 years) Untreated diabetes Female History of bilateral craniotomies & aneurysm treatment Note: WMS-III=Wechsler Memory Scale, Third Edition; LM = Logical Memory
48 was chosen instead of overall tremor se verity because the study hypotheses were based on upper tremor severity rather than tremor from other sources (e.g., vocal tremor would factor into a total TRS score but would lack a feasible mechanism for improving spatial ability). The median sco re was calculated and used as the dividing line to form Â“severe ETÂ” and Â“mild ETÂ” subgroups Figure 4-2 shows the distributions of the upper extremity tremor scores for the two subgroups. Table 4-2 provides group statistics concerning sociodemog raphic and clinical characte ristics. First, one-way ANOVAs and chi-square tests were perform ed to compare the ET groups and controls on sociodemographic variables. There were no group differences in age, gender ratio, education, handedness or the Barona full-scale IQ estimate, p s > .2. Next, the two ET groups alone were compared on indices of dis ease severity (t-tests ) to test to what extent the Â“mildÂ” vs. Â“severeÂ” distinction held across multiple measures of disease severity. Several significant differences we re found in the expected direction, beyond the total upper extremity tremor variable upon which the two groups were defined, t (16.23) = -5.72, p < .001, r = .82; severe ET group > mild group. Specifically, the severe ET group scored higher than the m ild ET group on measures of upper right extremity tremor, t (15.95) = -5.45, p < .001, r = .81, and upper left ex tremity tremor, t (17.31) = -4.80, p < .001, r = .76, axial tremor, t (20.26) = -2.50, p < .05, r = .49), the total TRS score, t (15.22) = -6.33, p < .001, r = .85, and ADL impairment, t (21.73) = -6.15, p < .05, r = .85. The severe and mild ET groups did not differ by age of tremor onset or the durat ion of tremor ( p s > .2). Individuals from the mild ET group, however, were diagnosed with ET more recently t han those from the severe ET group, t (20.54) = -2.54, p < .05, r = .49. The proportion of severe ET and mild ET participants on tremor-
49 Figure 4-2. Upper extremity Tremor Rating Scale (TRS) score distributions for the mild essential tremor (ET) and severe ET groups. Range: 0-20. Table 4-2. Demographic and ot her characteristics by group Demographic characteristic C ontrol ET mild ET severe N 33 16 15 Sex ratio, Female:Male 18:15 9:7 6:9 Age, years 71.3 (7.5) 73.1 (9.5) 69.8 (7.6) Education, years 15.6 (2.7 ) 15.4 (3.0) 14.3 (2.3) Barona Full-Scale IQ Estimate4 111.8 (6.3) 109.5 (7.0) 108.8 (5.6) Handedness, Right:Left 30:3 14:2 13:2 ET-specific characteristic Control ET mild ET severe Years since onset of tremor -15.3 (16.3) 19.9 (13.5) Years since diagnosis -5.3 (4.4) 11.6 (8.6)* Age at onset (years) -57.8 (18.6) 49.9 (17.1) TRS, Total Tremora -16.1 (6.0) 48.6 (17.3)* TRS, Axial Tremora -0.6 (1.3) 2.3 (2.3)* TRS, Upper Extremity Tremora,b -4.5 (1.6) 11.3 (4.2)* TRS, ADL Impairmenta -5.5 (3.2) 15.1 (5.0)* Tremor Medication (Yes:No) n/a 9:7 11:4 DBS History (Yes:No) n/a 0:16 3:12 Note: Values expressed in means (SD) unless otherwise indicated. TRS = Tremor Rating Scale, ADL = activities of daily living, DBS = Deep Brain Stimulation. a Score obtained in the "on" m edication state (if applicable). b Average score for left and right hands Severe ET group score > Mild ET group score ( p< .05). 4 Barona, Reynolds, & Chastain (1984)
50 reducing medications did not differ, 2(1) = 0.99, p = .32. Three memb ers of the severe ET group underwent DBS surger y compared with none in the m ild ET group. In sum, the controls and the severe and mild ET groups were matched on demographic (age, gender ratio, education, FSIQ estimate, handedne ss) and mood variables. The severe ET group scored higher on all tremor severity measures relative to the mild ET group. Fronto-Executive Functioning Tests Overview and Rationale After the initial screening and data collection, all participants completed a subset of tasks that have been associated with functioning of frontal-executive systems. These included tasks of attention/working memory cognitive inhibition, and speeded word generation. There were several reasons fo r administering these tasks. First, analyses were planned to determine whether these types of Â“fronto-executive functioningÂ” might be associated with performance on visuospatial tasks. Second, a variety of studies have consistently described mild deficits in fronto-executive functioning tasks in ET. As such, a similar profile would be expected in t he current sample of ET patients. Finally, the tasks used in the present study were also selected specifically because they did not require a hand-based motor response (which would be susceptible to upper-extremity tremor-related interference). The tasks are described in more detail below. Tests of Working Memory / Attention Â“Digit SpanÂ” subtest This subtest of the Wechsler Memory Scale-Third Edition (WMS-III; Wechsler, 1997b) is a measure of auditory attention and m ental manipulation. Strings of numbers are read by the examiner in increasing length. Participant s were required to repeat the strings exactly as they were presented (atte ntion condition), and in a second section, to
51 repeat them in backward order relative to how they were presented (manipulation condition). Performance was converted to standardized scores using comparison to a normative sample matched by age. Â“Spatial SpanÂ” subtest This subtest of the WMS-III is a visual analog of the Digit Span subtest. The experiment uses a finger to touch increas ingly long sequences of blue cubes (ten of which are affixed in a roughly evenly dist ributed arrangement over a white board). Participants must touch the same sequence of blocks in the same order as the experimenter (attention condi tion), and then in a subsequent section, in the reverse order as they were presented (manipulati on condition). Again, age-matched norms from the WMS-III scoring manual were used to convert scores to a standardized metric (Wechsler, 1997b). Test of Inhibition: Â“Str oop Color and WordÂ” Test On the Stroop Color and Word (Stroop CW ) test (Golden, 1978) one must name the colors of ink in which a list of words is printed (i.e., Â“redÂ”, Â“b lueÂ”, and Â“greenÂ”). The words themselves are the names of colors (i .e., Â“redÂ”, Â“blueÂ”, and Â“g reenÂ”). Because the color and word never match for any given item participants must inhibit the relatively automatic behavior of reading the words them selves in order to correctly name the incongruent color of the ink in which the word is printed. Performance is measured in terms of the number of correct ly named colors in 45 seconds (Interference condition). Two other Â“baselineÂ” trials are administered as well for co mparison to performance in the interference condition: in the first, participants simply read a list of color names printed in black, and in the second, they must name the different colors of ink in a series of printed Â“XXXXÂ”s. Age corrections are app lied to the raw scores. The corrected raw
52 score for the interference trial (administered third) was used as a measure of Â“inhibitionÂ” in this study. Tests of Verbal Fluency: Â“Letter Fl uencyÂ” & Â“Category FluencyÂ” Tests A letter fluency test and a category fluency te st were used in this study to provide measures of verbal fluency / word generation using lexical/letter (F A, S) and semantic (animals) rules for word generation, as de scribed by Spreen and Benton (1977). Verbal output over one minute is measured for eac h condition, providing a total score representing letter fluency and a total score representing category fluency. Standardized scores are calculated from t hese scores using demographically corrected norms (Heaton, Miller, Taylor, & Grant, 2004). Visual Cognition Tests Overview & Rationale A battery of clinical neuropsychologica l tests and computerized experimental tasks was constructed to test two subdomains of vi sual cognition in ET: spatial cognition and object/form cognition. The Â“spatialÂ” tasks included one commonly used measure in clinical settings (Judgment of Line Orientat ion) and three experimental tasks: Visual Search Task, Mental Rotations Task, and Computer-Generated Arena. The Â“object/formÂ” tasks again included one commonl y used clinical meas ure (Benton Facial Recognition Test) and one experimental task (Object Attention Task). All the experimental tasks were administered on a computer. These te sts are described in more detail in the sections that follow.
53 Clinical Neuropsychological Tests Â“Judgment of Line Orient ationÂ” (JOLO) test BentonÂ’s Judgment of Line Orientation (J OLO) test (Benton, Hamsher, Varney, and Spreen, 1983) requires participants to identif y the angular orientati ons of lines using a reference array of 11 different radii arrange d in a semicircle. Each of 30 items are worth one point each and contain two lines whos e spatial relationships are to be judged (the point is lost if the answe r is incorrect for either line) The JOLO was administered for two purposes: to replicate the previous finding of high spatial performance in most of the severe, presurgical ET patients tested in a previous study (Springer et al., 2006) and to provide a clinical measure of spatia l perception to be used as evidence for or against a Â“visuospatial superiority effectÂ”. Raw scores were converted into percentiles relative to a normative population, after adjustment for gender and age (Benton, Sivan, Hamsher, Varney, & Spreen, 1994). Â“Facial Recognition TestÂ” (FRT) On the Facial Recognition Test (FRT ; Benton, Sivan, Hamsher, Varney, & Spreen, 1994), participants must match a target individual (photograph of a face) to faces in an array of six, similar-looking peop le. There are three conditions, whereby the subject must (1) select one face in the array that matches identically to the target face, (2) select three faces shown in front and thr ee-quarters views that match the identity of the target individual, and (3) select three faces shown under unusual lighting conditions that match the identity of t he target individual. Answ ers are reported verbally, referencing photos by number, or by pointing. There are 27 trials on the short form, and scores range from 0 to 27. This raw score is converted to a long-form equivalent score,
54 which is then adjusted for age and years of education completed (Benton, Van Allen, Hamsher, & Levin, 1978). Computerized Experimental Tasks Common physical setup across tasks Each of the three experimental tasks was administered on a laptop computer (Intel CoreÂ™2 Duo CPU T7250 at 2.00 GHz, 2.00 GB RAM). Stimuli were displayed on a 14Â” monitor with a 1280x800 pixel resolu tion. While computerized testing in the elderly may raise concerns about whether familiarity with computers may influence cognitive testing, these pot ential concerns were thought to be outweighed by several advantages: (1) the managemen t of stimuli presentation using precise timing parameters; (2) the ability to incorporate dynamic (and/or more ecologically valid) visuospatial stimuli that ar e impossible to present using paper-and-pencil-type tests (e.g., a changing, first-person perspective of a 3-D virtual room); (3) the ability to precisely and reliably measure response ti me data; and (4) efficient computation and management of participant data with secu re, password-protec ted storage. Responses were recorded using one of tw o input devices, depending on the task: a handheld videogame control pad, or a customized Â“button boxÂ”. Both are depicted in Figure 4-3. The navigation pad was held in both hands (not affixed to the table or computer) and operated by the le ft thumb. During C-G Arena p ilot testing, it was found that ET patients were able to navigat e with greater ease using the control pad compared with a joystick control or the arro w keys on the keyboard. The interfering effects of tremor on the use of the game controls were minimized because the entire navigation pad moved synchronously with any tr emor, and thumb-relat ed tremor tended
55 (A) (B) Figure 4-3. Input modalities for the computerized tasks. (A) A Â“button boxÂ” was used for four of the experim ental tasks. The two Â“targetÂ” buttons with the hand positioning is shown. (B) A handhel d, gaming navigation pad was used to operate the Computer-G enerated Arena, with nav igation controlled entirely by the left thumb as shown (the other buttons had no functionality).
56 to be minimal. For these reasons, t he handheld videogame navigation pad was used for the C-G Arena Task in place of the standard joystick control or arrow keys. The button box consisted of two large, plastic circles (diameter approximately three inches) mounted to a black tray that wa s overlaid on the computer keyboard. The right button was dull green in color, and the left button was white. This color coding allowed for precise differentiation between buttons. All subjects responded with their right hand on the right button, and left hand on t he left button. The button box was used to record accuracy (i .e., correct, incorrect) and res ponse time (seconds, with accuracy to the hundredths place) in the following co mputer-based tasks: Mental Rotations Task, Visual Search Task, Object Attention Ta sk, and a Choice Reaction Time task. For these four tasks, stimulus presentation and response collection were controlled by customized software written in Visual Basic 6.0. Throughout testing, participants were allowed to sit in such a way that they could reduce the amplitude of any tremor throughout the experi ment (usually, with their forearms resting on armrests and the edge of the table). Based on pilot testing, ET patients with mild to moderate upper extr emity tremor appeared to accurately and efficiently respond using the button box or the navigation pad. Even so, an additional attempt was undertaken to empirically assess each participantÂ’s efficiency in using motor controls. This was done via a Choice Reaction Time task using the button box and via a motor control task using the navigati on control pad. Each participant was also interviewed and provided ratings regarding diffi culty of use and comfort with the controls during computerized testing.
57 Â“Choice Reaction TimeÂ” (CRT) task Rationale. This task was used as a basic Â“mot or speedÂ” task to test potential group differences in simple motor speed when using the manual computer buttons (i.e., mild ET vs. severe ET vs. normal controls in leftand right-handed visuomotor reaction time). At least one previous study (Ku mru, Begeman, Tolosa, & Valls-Sole, 2007) found no differences in simple manual RTs between elderly ET patients and elderly controls. It was important to determine wh ether a similar effect existed in the present sample, as any significant RT differences found between groups could impact interpretation of the RT data extracted from the ta sks that used this input device (i.e., the Visual Search Task, Mental Rotations Task, and Object Attention Task). Stimuli and procedure. Two stimuli were used on the CRT: the word Â“RIGHTÂ” with a large white arrow below it pointing ri ght, and the word Â“LEFTÂ” with a large white arrow below it pointing left. The stimuli are shown in Figur e 4-4. The word and arrow were presented simultaneously for each stim ulus. Each of the two stimuli was presented 10 times, intermixed in one of two pseudorandom orders, for 20 total trials. Participants were instructed to press the le ft button of the respons e box with their left hand as soon as they recognized the word Â“L EFTÂ”, or the right button with their right hand as soon as they recognized the wo rd Â“RIGHTÂ”. Each trial began with the appearance of the word/arrow stimulus, locat ed centrally on the computer screen. The stimulus remained visible until the participant responded with a left-button press (using the left hand) or a right-butt on press (using the right han d). Response times were recorded to the nearest hundredth of a second and recorded Â“correctÂ” if the stimulus direction matched the side of the response. Mismatches were coded as Â“incorrect.Â” The next trial began after a random interval of 2-4 seconds.
58 (A) (B) Figure 4-4. Choice Reaction Time (CRT) task st imuli. (A) The Â“leftÂ” stimulus, requiring a speeded button press with the left hand on the left button. (B) The Â“rightÂ” stimulus, requiring a speeded button press with the right hand on the right button.
59 Data reduction. Median RTs were computed separately for the left and right hands over correct trials (Ratcliffe, 1993). In correct trial data were not used. Thus, two dependent variables were extracted for each participant: Â“Left RT Â” and Â“Right RTÂ”. Â“Mental Rotation TaskÂ” (MRT) Description. The paradigm used in the present study was modeled after Funk and Brugger (2007) and used single photos of hands rotated in 2-dimensions (i.e., with the plane of the page, like the spinning of a needle in a navi gational compass). When using and viewing oneÂ’s own hands to intera ct with objects skillfully in the nearby environment, with or without visual guidance, spatial abilities und ergo use. Potentially, this may amount to a measurable Â“practice effectÂ” of spatial elements of visual cognition specifically related to the hands. This practi ce effect would occur with hands to a higher degree than with objects because greater experience with hands as stimuli increase the ease with which mental representations of hands can be created (Bethell-Fox and Shepard, 1988). Moreover, as ET most often affects the intentional movement of the hands and upper extremities, this ability may be practiced more often particularly in this population. For these reasons, hand stimuli were incorporated in this mental rotations task. Stimuli. Stimuli were created using two diffe rent actors: one male and one female. For each actor, one picture of the right hand in the dorsal view (fingers outstretched) was taken to create two core stimuli (one male right hand and one female right hand), from which the rest of the stimuli were derived. The photos were taken in black and white under similar lig hting conditions and backgrounds (a featureless wall). First, left versions of t he right male and female hands were created by horizontally flipping them in a digital image editor. Each of these four stimuli (male-right, male-left,
60 female-right, female-left) were used to create sets of fo ur stimuli representing four angles of rotation: fingers pointing to t he top of the image (0 degrees), rotated 90 degrees to the right, upside down (rotated 180 degrees), and rotated 270 degrees to the left. Only these 4 orientations were used in order to keep test ti me to a minimum. Thus, the final stimulus set consisted of 16 unique hand stimuli: 2 actors (male/female) X 2 hands (right, left) X 4 positions (0, 90, 180, 270 degrees), and the eight stimuli from each actor were indistinguishabl e from each other. Two exam ples of these stimuli are provided in Figure 4-5. Procedure. The Mental Rotation Task (MRT) was administered on the laptop computer, using the response button box to record the timing and accuracy of responses. The participantÂ’s task was to decide as quickly as possible whether the displayed stimulus represented a Â“leftÂ” hand or a Â“rightÂ” hand. The following instructions were read aloud by the examiner, while a wri tten set of instructions was displayed on the computer screen: Â“ In this task, you will see a seri es of photos, one at a time. Decide whether it is a left hand or a right hand. If it is a left hand, then press this button with your left hand (examiner points to the left but ton). If it is a right hand, then press this button with your right hand (examiner poi nts to the right button). Respond as quickly and as accurately as you can. LetÂ’s do a few trials for practice before trying the test, just so you get the hang of it. Â” Participants were administered 16 practice trials. They received feedback on a trial by trial basis regarding whether the response was correct. This was provided by the examiner and via a visual text box on the computer monitor. This practice section was used to ensure that the participants underst ood the task instructions. Following this
61 (A) (B) Figure 4-5. Mental Rotation Task (MRT) exampl e stimuli. (A) A Â“leftÂ” hand stimulus rotated 90 degrees inward from vertical (0 degrees), requiring a Â“leftÂ” hand button response. (B) Another Â“leftÂ” hand stimulus, rotated at 90 degrees outward, also requiring a Â“leftÂ” button response.
62 practice section, the exam iner told the participant, Â“ Do you have any questions? Now you have the hang of it. You will now be present ed with several sets of items, but you wonÂ’t be told whether you are right or wr ong. You will be provided with occasional breaks. Remember to respond as qui ckly and as accurately as you can .Â” Following these instructions, the test trials comm enced. No feedback was provided concerning errors or correct responses during the test trials. Each trial of the MRT began with a fixa tion mark, a crosshairs (Â“+Â”), that appeared for 500 ms in the center of the screen. This was immediately followed by a hand stimulus which remained on the screen until the subject responded. Participants were given as long as they wished to respond. There was an inter-t rial interval of 2 seconds before the onset of a new fixation ma rk for the subsequent tr ial. A total of 64 test trials were administered and consisted of 4 blocks of 16 trials each with the primary stipulation of this order bei ng that no stimulus orientati on could be presented more than 3 times in succession. Each block cons isted of a unique psuedor andomized order of the 16 photos in the stimulus set. After co mpleting each block of stimuli, a message box appeared on the screen asking the partici pant whether a brief break was needed. Thus, opportunities for breaks we re provided prior to the 17th, 33rd, and 49th trials, with most participants choosing to continue the task within about 10 seconds. The total duration of this task was 10-15 minutes. Data Reduction. Median response times (Ratcli ffe, 1993) and error rates were obtained for each angle of rotation. Four angl es of rotation were defined for the left and right hand stimuli: (1) 0-degrees (upright), (2) 90-degr ees rotated Â“inwardÂ” (90-degrees rotated clockwise for the left hand, and 90degrees rotated counterclockwise for the
63 right hand), (3) 90-degrees rotated Â“outwardÂ” (90-degrees rotated counterclockwise for the left hand, and 90-degrees rotated clockwis e for the right hand), and (4) 180 degrees (upside down). These conditions were abbrev iated as Â“0Â”, Â“90-in Â”, Â“90-outÂ”, and Â“180Â”, respectively5. The left and right hand stimuli, grouped by these four conditions, are depicted in Figure 4-6. Â“Visual Search TaskÂ” (VST) Rationale. The Visual Search Task (VST) wa s administered using the laptop computer with the button box to record responses. The task was adopted from the Â“hard feature serial visual searchÂ” task of Elli son, Schindler, Pattison, and Milner (2004). It requires participants to determine whether a particular, predefined target exists in a field of similar-looking, non-ta rget items (using noncovert, free-ranging gaze). In their study of healthy adults, Ellison et al. found t hat the application of repetitive transcranial magnetic stimulation (rTMS) to the right super ior temporal gyrus (rSTG) was associated with impaired performance on the VST (slowe r response times), but not on another task typically used in the diagnosis of neglect. Admi nistration of rTMS to the right posterior parietal cortex (rPPC) showed the opposit e pattern. Another research group (Gharabaghi, Berger, Tatagiba, & Karnath, 2006) supplemented these findings by demonstrating chance performanc e (51.3% correct) on this task in a patient during intraoperative, direct cortical st imulation of the middle portion of the rSTG (stimulation of this area did not affect performance on a simple feature visual search task or on a hard 5 Pairing the left and right hand stimuli in this manner was done because the efficiency of mental rotating hand stimuli is influenced by biomechanical constraints; t hat is, it is difficult to mentally imagine and rotate images of hands that are physically difficult to generate with oneÂ’s own body (e.g., the dorsal view of oneÂ’s own upside-down hand) (Funk & Brugger, 2007). From this viewpoint, left and right hand stimuli are associated with different difficulty levels if they are paired by absolute angle of clockwise rotation, i.e., both hands rotated 90-degrees clockwise, or both ha nds rotated 90-degrees, counterclockwise. In contrast, leftand right-hand stimuli are similar in difficulty if Â“90-degrees rotated Â‘inÂ’Â” and Â“90-degrees rotated Â‘outÂ’Â” groupings are used.
64 (A) (B) (C) (D) Figure 4-6. MRT example st imuli grouped by increasing degree of Â“biomechanical constraint.Â” Left and right hand stimuli are shown at the following angles of rotation: (A) 0 degrees, (B) 90 degrees Â“i nwardÂ”, (C) 90 degrees Â“outwardÂ”, (D) 180 degrees (i.e., abbreviated 0, 90-in, 90-out, and 180). The female version of the hand stimulus set is shown.
65 conjunction search task). The patient had performed at 86. 7% correct during an intraoperative baseline condition, and close to 100% performance prior to and after surgery. Direct stimulation of adjacent cortical areas, including the rPPC, demonstrated no significant reductions in performance. Th ese two studies provid e evidence that the rSTG is critical to hard-featur e visual exploratory search and identification, at least, as measured by this task. The volumetrics st udy of ET by Daniels et al. (2006) found evidence of increased cortical density in ri ght superior temporal gyrus (in patients with more longstanding ET). Thus, the visual search task of Ellison et al. (2004) was selected due to its potential for providing a sensitive means for detecting the presence of group differences, at least in this form of visuospatial processing (serial, hard-feature, exploratory visual search). Stimuli. In the VST, participants indicated the pr esence of a unique target Â‘LÂ’ in a field of non-target items. T he non-target items were identica l to the target stimuli (i.e., shape, size, color), but differed in terms of or ientations (rotated clockwise at 180 or 270 degrees). Eighty trials were cr eated, each containing arrays of these items. Half of the trials contained one unique target Â‘LÂ’ with se ven non-target items. The remaining 40 non-target trials contained eight non-target items. An equal number of each type of non-target item (180 or 270 degrees rotated Â‘LÂ’) was replaced by the target in non-target trials. In each trial, the eight items were distributed pseudorandomly and relatively evenly across the screen, with 2 items in each quadrant (upper left, upper right, lower left, lower right). Among the 40 target tr ials, the target appeared in each spatial quadrant an equal number of ti mes (10), and its appearance in a given quadrant was balanced between the first and second halves of the test (5 each). Targets also
66 appeared equally in the periphery (leftmost 25% of the scr een, rightmost 25% of the screen) and the central area (middle 50% of t he screen). Two stimulus orders were created and distributed across parti cipants, with neither Â“non-targ etÂ” nor Â“targetÂ” trials exceeding three in succession. Figure 4-7 s hows example stimuli from the non-target and target conditions. Procedure. This task was administered on t he laptop computer and subjects responded manually using a button box. The fo llowing instructions were read aloud by the examiner, while a written set of instructions was disp layed on the computer screen: Â“In this task, you are to look for a particu lar shape among a set of shapes spread across the screen. All the shapes will look similar, but this is the one you are looking for (examiner points to the target stimulus on the screen). If you see it on the screen, press this button with your right hand (examiner points to the right button). If not, press this button with your left hand (examiner points to the left button). Respond as quickly and as accurately as you can. LetÂ’s do a few tria ls for practice before tr ying the test, just so you get the hang of it.Â” Participants were administered 10 practice trials, with visual (message box) and auditory feedback (examine r) as to whether each response was correct. This practice section was used to ensure that the participants understood the task instructions, and additional instructi ons were provided until the participants understood how to respond. Following this pr actice section, the examiner told the participant, Â“Do you have any questions? Now you have the hang of it You will now be presented with several sets of items, but y ou wonÂ’t be told whether you are right or wrong. You will be provided with occasional breaks. Remember to respond as quickly and as accurately as you can.Â” Following thes e instructions, the test trials commenced.
67 (A) (B) Figure 4-7. Visual Search Task (VST) exam ple stimuli. (A) A Â‘non-targetÂ’ stimulus array, requiring a Â“No/LeftÂ” button res ponse. (B) Example of a stimulus array containing the target Â“LÂ”, r equiring a Â“Yes/RightÂ” button response.
68 Across practice and test sections, each tr ial was preceded by a central fixation cross, Â“+Â”, that appeared for 500 ms, immedi ately followed by presentation of the stimulus array. Each stimulus array re mained present until the participant pressed one or the other button, at wh ich point the array disappear ed for 2000 ms (black screen) prior to the appearance of the ne xt fixation point. After ev ery 20 trials, a message box appeared on the screen asking the participant w hether a brief break was needed. Thus, opportunities for breaks were provided prior to the 21st, 41st, and 61st trials, with most participants choosing to continue the task within about 10 seconds. The total duration of this task was 10-15 minutes. Data reduction. The two major dependent variabl es included response time (seconds) and error rates. These were derived from each of the two experimental conditions: a) trials on which a target was present and b) trials without a target stimulus. Median response times were derived fr om the correct trials (Ratcliffe, 1993). Error rates over all trials per condition we re expressed as a value between 0 and 1. Thus, for each participant, four dependent variabl es were extracted: Â“Target RTÂ”, Â“Nontarget RTÂ”, Â“Target Error RateÂ”, and Â“Non-target Error RateÂ”. Â“Computer-Generated ArenaÂ” (C-G Arena) task Overview. The Computer-Generated Arena (C-G Arena) task is a computerbased spatial navigation and memory task m odeled after the Morris Water Maze, a paradigm classically used for studying spatia l navigation and memory in rats (Jacobs, Laurance, & Thomas, 1997; Jacobs, Thomas Laurance, & Nadel, 1998; Thomas, Hsu, Laurance, Nadel, & Jacobs, 2001; http://web .arizona.edu/~arg/). The participants viewed a computer monitor upon which they re ceive a first-person view of a virtual, 3-D room. A handheld, thumb-contro lled direction pad is used to navigate within the room:
69 pressing up/down and left/right on the direction pad changes the first-person view on the monitor, such that thes e inputs simulate forward/backw ard movement, or spinning left/right. Navigation is constrained to a circ ular Â“arenaÂ” designed to look like a pool of water. The arena is bordered by a low wall and is in its entir ety located in a perfectly square room, with each of the f our walls of the room differ entiated from the others by 12 unique, fractal-based patterns. Figure 4-8 depicts a schematic of the relative spatial arrangement of the roomÂ’s elements, as well as views of each of the walls. Procedure. First, each participant was provided wi th up to 10 minutes of practice with the controller for navigati on within a Â“practice roomÂ”. The room contained a visible white target and distinct, uniformly colored walls. The experimente r first modeled the use of the controller to move forward/ backward (up/down on the control pad), spin left/right (rotation without translation; left/right on the control pad), and then combining directions using diagonal movements to mo ve in arcs (i.e., trans lating forward/backward while spinning left/right: forw ard-left, forward-right, backwar d-left, backward-right). The participant then practiced these movements unaided until a general level of proficiency and ease of use was observed and reported. Following the practice session and demonstration of adequate master y of the navigation controls the test trials of C-G Arena commenced. Apart from this practice period (no participant took more than 5 minutes), the test trials of C-G Arena were divided into two sections/conditions, a Â“visible target trialsÂ” section and a Â“hidden target trialsÂ” section. C-G Arena Â“visible target tr ialsÂ” (motor control task). This section of C-G Arena was designed as a motor control task, administered to provide measures of simple motor adeptness using the controls to navigate as intended. Additionally, it
70 (A) (B) (C) (D) (E) (F) Figure 4-8. Computer-Generated Arena (C-G Arena) example views as seen by participants. Pictured fr om the Â“hidden target tr ialsÂ” (memory/learning condition) are the (A) northwest quadrant, (B) northeast quadrant, (C) southwest quadrant, and (D) southeast quadrant. (E) A view of the northwest quadrant containing the targe t, shown after targ et acquisition. (F) A view of the C-G Ar ena from the Â“visible tar gets trialsÂ” (mo tor control condition).
71 yielded data allowing analyses to rule out any motivational or perceptional problems when interacting with this virtual program. Group differences on the measures of this task would have implications upon the main C-G Arena spatial memory variables. For example, a potential group difference could be masked if ET patients possessed better spatial learning/memory for differences in basic motor control that would impact interpretation of results dependent the hidden ta rgetÂ’s location but, because of tremor, they had a more difficult time in using the controls in navigating as desired. On this Â“visible target trialsÂ” secti on of C-G Arena, participants had to navigate within a room from different starting points and orientations (the init ial direction faced) over 8 trials, with the objective of locating a vi sible, elliptical white target on the floor (also in different locations across trials) as quickly as possible. The target was within the field of view from the beginn ing of the first four trials, but on the last four trials, the participants had to turn substantially to bring the target within the field of view. The target was Â“locatedÂ” when the participant successfully used the controls to navigate to the target and Â“standÂ” on top of it. Participants received immediate auditory feedback for a successful trial with presentation of a looped sound clip (i.e., Â“ta daaÂ”). All participants received the same pseudorandomized order of start posi tions and target locations across the 8 trials (specified in Appendix B). Two dependent variables were extracted from this set of tria ls: (1) the length of the paths taken from starting positions to the visible targets (i.e., Â“tot al path length of visible trialsÂ”; summed over the 8 trials), and (2) the time taken from the ons et of the trial to reach the targets (i.e., Â“total latency of visible trialsÂ”; also summed over the 8 trials). No parti cipant demonstrated such poor
72 control of the navigation device (due to tremor or other reasons) th at testing could not proceed (subjective judgment by the experimenter). C-G Arena Â“hidden target trialsÂ” (spatial navigation and memory). The main dependent measures of spatial navigation and memory were ex tracted from this section of the C-G Arena task (Â“hidden ta rget trialsÂ”). Participants were required to find a hidden, spatially fixed target on the floor of a new room ov er 8 trials. All participants received the same pseudorandomized order of start positions and in itial orientations across trials, as specified in Appendix B. Fi rst, they were provided a fixed Â“tourÂ” of the new room, whereby the experim enter started from the sout h wall, oriented north, and navigated to the center of the room. The experimenter faced the north wall for 2 seconds before turning to the right to face the east wall. Turning clockwise in this manner, each wall was displayed for 2 seconds until the north wall was displayed again. The experimenter then turned counterclo ckwise and again faced each wall for 2 seconds before facing the north wall. During this Â“tour,Â” the expe rimenter mentioned to each participant that a hidden ta rget will be placed in a fixed location inside this room, and that they would have to search this room to find it as quickly as possible over each trial. Participants were informed that findi ng the target would be associated with visual feedback (the target appearing in white) and auditory feedback (presentation of the same looped Â“ta daÂ” sound clip as was heard dur ing the visible target trials. Finally, participants were told that failure to find t he hidden target after 120 seconds in the first two trials would result in the experimenter taking over the controls at 120 seconds and finding the target for them (but that they should try their best to find the target on their
73 own). Thus, target acquisition was guaranteed for all participants in the first 2 trials. No help or additional instructions were provided after the first 2 trials. In this Â“hidden target trialsÂ” section of C-G Arena, spatial navigation and memory were assessed by (1) Â“total path lengthÂ” over 8 trials to reach the ta rget or the end of the trial (120 seconds), whichever occurred first, (2 ) Â“total latencyÂ” (s econds) over 8 trials between trial onset and reaching the target or the end of the trial, whichever occurred first, and (3) Â“total number of targets foundÂ” over 8 trials. After the 8th trial, the hidden target was removed from its fi xed location (covertly to the par ticipant) in order to ensure that target acquisitions were not simply a result of trial and e rror. In this last (ninth) Â“probeÂ” trial, spatial memory was assessed with the percent of time spent over the course of the trial in the co rrect quadrant of the room (i.e., where the target was located during the first 8 trials). The fourth measure of spatia l learning and memory from the hidden targets trials, then, was (4) Â“percent time in target quadrant (probe trial)Â”. This probe trial was important to ensure that participants were using the spatial cues to locate the target. If they we re, they would spend the majority of the probe trial within the area of the arena where the platform was loca ted during the initial eight trials. Figure 4-8 shows depictions of this room used for the hidden targets trials, while Figure 4-9 shows a visual schematic to hel p illustrate how some of the dependent variables were calculated on the C-G Arena task. Appendix B pr ovides the images used to texture the roomÂ’s obj ects as well as the specific task parameters used in C-G Arena, for both the visible targets condi tion and the hidden ta rgets condition. Â“Object Attention TaskÂ” (OAT) Overview. This experimental task provided a measure of objectbased attention and was therefore used as a Â“controlÂ” task, as severe ET patients were expected to
74 (A) (B) (C) Figure 4-9. C-G Arena example path plots depicting spatial navigation and memory acquisition (Â“hidden target trialsÂ”). Plots depict the sample performance from a young female pilot participant, showing a birdÂ’s eye schematic of the arena (available to the examiner only). Yellow arrows (which indicate start positions/orientations), green lines (which indicate paths taken), and red XÂ’s (which indicate the end point of navigation) are shown, with the small white circle indicating the position of the hidden target. (A) Navigation is based on trial and error stra tegy for the first trial because the target is hidden and the location is not initially known. The target becomes visible when the participant passes over it. (B) Trial 7: after the fixed target location is found and lear ning occurs, subsequent trials such as this one are associated with more di rect paths to the location of the hidden target (reflected by reduced pat h lengths and shorter latencies). (c) Probe (9th) trial: the target is removed, and participants who have learned the target location well spend mo st of the trial time searching the quadrant of the arena previous ly containing the target. This is reflected by a high percentage of the time spent during the probe trial in the Â“northwestÂ” quadrant.
75 outperform mild ET patients only on spatially based tasks. This Â“object attention taskÂ” (OAT) was adapted from Behrman, Zemel, & Mozer (1998), who used this task in young, healthy adults to help provide evid ence for a distinction between object and spatial attention strategies in visual percept ual processing. In their study, participants had to attend to and make simple judgments regarding two visual Â“featuresÂ” apparent on overlapping rectangles. This task had two main conditions: (1) the two features were located on opposite ends of one of the rectangles (the one overlapping or overlapped by the other), or (2) on the two different rectangles. In the latter condition, the features in question are located closer together (see Figure 4-10 for examples of the stimuli). The use of a purely spatial attent ion strategy would pr edict faster reaction times for the simple judgments when the feat ures were located on the two separate objects, because the features are closer t ogether than in the single-object condition. Yet, the authors found the opposite effect: reacti on times were faster when the features were located on opposite ends of the single ob ject (spatially further apart), suggesting that object-based attention strategies are us ed in this task (for healthy and relatively young individuals). Object-based attention phenomena such as this have been observed by others examining object-based, spatia lly invariant shifts of attention or maintenance of attention (e.g., OÂ’Craven, Downing, Kanwis her, 1999; Yantis & Serences, 2003). These studies have lent further empirical support fo r object-based attention as a construct that is separate from spatial attention. T hey also partly have elucidated their neural networks using functional neuroim aging technologies such as fMRI.
76 (A) (B) (C) (D) (E) (F) Figure 4-10. Object Attent ion Task (OAT) example stimuli. Top row: examples of stimuli requiring a Â“same objectÂ” judgment in the (A) one-object (unoccluded), (B) two-object, and (C ) one-object (occluded) conditions, respectively. Bottom row: examples of stimuli requiring a Â“different objectÂ” judgment, in (D) one-object (unoccluded), (E) two-object, and (F) one-object (occluded) conditions, respecti vely. Use of primarily spatial attention would predict slower res ponse times for same and different judgments in the one-object conditions, because in this condition, the features to be judged are spatially farther apart t han in either of the oneobject conditions (the distance is 41% greater). In healthy young adults, however, the fastest response times occur in the one-object conditions, suggesting that the normal response on this task utilizes object-based attention (Behrman et al., 1998).
77 Stimuli. As stated above, Figure 4-10 shows examples of t he task stimuli. They were presented in black over a white background and were composed of two overlapping rectangles in an Â“XÂ” pattern, with one rectangle overla pping the other. Two of the X-shaped figureÂ’s four Â“armsÂ” contained a set of features (Â“bumpsÂ”), made up of either two or three divisions at the edge. The number of bumps on each of the two arms was either the same (2-2 or 3-3), or different (2-3). These features appeared in two possible conditions: (1) at the ends of the same rectangle (the overlapping rectangle, or the rectangle that is overlapped by the other), (2) one feature on each of the two different overlapping rectangles (i.e., with the features spat ially located at the bottom, top, left, or right pair of ends). The overlapp ing rectangles were oriented diagonally from top-left to bot tom right, or diagonally bottom left to upper right in the figure. The entire stimulus set comprised ev ery possible feature pa iring and their spatial configurations, yielding 32 total stimu li. These stimuli and subcategories were constructed to match the specificati ons outlined by Behrman et al. (1998). Task. The experimental procedure for this ta sk also was modeled after the study by Behrman et al (1998). In each trial, a black fixation dot appeared centrally on the computer screen for 500 ms. This was fo llowed by a delay of 1000 ms with a blank screen, and then a stimulus from the set. The par ticipant was told to view the figure and decide, from the two edges of the figure containing bump s, whether the number of bumps on one edge was the same or different from the number of features on the other edge. Responses were indicated with the tw o available buttons on the response board, and each participant was told to respond as qui ckly and accurately as possible. Each
78 response coincided with the stimulus disapp earing from the screen for 1 second before the next trial. The order of stimuli presented to eac h subject consisted of 4 blocks of pseudorandomized trials. These blocks were separated by brief pauses with durations as needed by the participants (typically less t han 30 seconds). Each block consisted of 32 trials, half of which required Â“sameÂ” resp onses, and the other half required Â“differentÂ” responses. Among the 16 Â“sameÂ” trials, there were an equal number of 2-2 trials and 33 trials, and the locations of these featur es were counterbalanced in their spatial distribution (i.e., located on bot h rectangles in the figure, or the same). The spatial location of the features among the 16 Â“diffe rentÂ” trials was also counterbalanced. Prior to initiating the task, t he administrator stated to each participant, Â“In this task, you will see a pair of rectangles that overlap to form an Â“XÂ” with four ar ms. Here are some examplesÂ” (instructor points to ex ample figures on the screen). Â“ You will notice that on two of the four arms, there are either one or two bumps on the end. Your task is to indicate if the number of bumps on one arm is the same or different from the number of bumps on the other arm. You would indicate Â“sameÂ” for this one (point) because the two arms with the bumps both have 2 bumps eac h. You would also say Â“sameÂ” for this one (point) because the two arms with the bumps both have 3 bumps each. So the number of bumps on each arm is the same, and so you would press this button (point). On the other hand, if the number of bum ps on one arm does not match the number of bumps on the other arm, then you would press this button (point) to say that they are different in number. So for this one you woul d press Â“differentÂ”, and for this one, you would press Â“differentÂ”, because one arm has 2 bumps, and the other has three. LetÂ’s
79 try a few practice trials so you get the hang of the directions. Respond as quickly and as accurately as you can.Â” Following these instructions, participants were provided with 16 trials and visual and auditory feedback about errors following each one. The experimenter confirmed t hat the instructions we re understood and answered any questions, and then the test trials commenc ed (no feedback about errors or correct responses). The total duration of this task was 10-15 minutes. Data reduction. The raw data were reduced into response times over correct trials and error rates. These were grouped by the following conditions: Â“one objectÂ” (object attention) and Â“two objectsÂ” (spatial att ention). Thus, the important point is that subjects were told to answer with Â“sameÂ” or Â“differentÂ” responses pertaining to whether the number of bumps matched; however, analyses distinguished only between whether the features that were judged were localized to one rectangle in the figure (one-object condition), or two different rectangles in the figure (two-objects condition). Data Reduction & Analysis Plan for the Study Reaction time (RT) variables were extracted from the re levant conditions of each of the timed computerized test s (Visual Search Test, Obje ct Attention Test, Mental Rotations Test, and Choice Reaction Time Test) for each participant group (mild ET, severe ET, and controls). Each participantÂ’s RT data were calculated using the median RT value of all non-excluded (correct) trials (Ratcliffe, 1993) per condition and task. Prior to each statistical analysis, all dependent variables were examined for normality, separately within each of the three subject groups. Log10(RT+1) transformations were applied to response time (RT) variables within a given analysis if Kolmogorov-Smirnov tests found any to be non-normally distributed ( p < .05). Most experimental task error rates were also found to be non-normal using this criterion, and
80 standard error rate transfo rmations (arcsin, arcsin x) were applied in these cases. Kolmogorov-Smirnov tests were re-performe d after any transforma tions to re-check normality ( p s > .05) before proceeding with param etric analyses. In cases whereby transformations insufficiently improved the no rmality of these distributions (KolmogorovSmirnov test p s < .05), non-parametric statistical pr ocedures were performed. In all cases, non -transformed means and standard deviation s are reported in the text, tables, and figures. The text specifies whether any transformations were applied prior to a given statistical analysis, as well as the type of analysis performed. The primary analytic approach involved analyses of variance (ANOVAs) for Specific Aim 1 and hierarchical regressions for Specific Aim 2. Specific aim 1 tested the hypothesis that severe ET has superior spatia l cognition skills vs. mild ET, relative to controls. The associated analyses were des igned to compare the three groups (severe ET, mild ET, control) in their performances on tasks of visual co gnition, first for the spatial tasks (JOLO, MRT, VST, C-G Arena) and then for the object/form tasks (FRT, OAT). These analyses used ANOVAs or their non-parametric equivalents, and deconstructed main effects or interactions with post-hoc contrasts. Adjustments were made when appropriate in these analyses, e. g., the use of corrections for inflated familywise error rates (multiple tests) or adjustments to degrees of freedom due to violations of sphericity or variance assumptions. Specific aim 2 tested the hypothesis that ET seve rity (i.e., upper extremity tremor severity) is indeed a factor t hat contributes to spatial func tioning, and not other potential explanatory variables such as education leve l, full-scale IQ, or age. Thus, it was predicted that tremor severi ty predicts a significant amount of variance in spatial
81 functioning, above and beyond other demographic, c ognitive, or mood predictors. Only the 31 ET patients were used in these analyses. Spearman correlations were performed between spatial outcome variables (those that sh owed the Â“spatial superiority effect in severe ETÂ”) and potential predict ors, which included upper extremity tremor severity, and demographic, cognitive (i.e., fron to-executive functioning measures), and mood variables. Variables correlating signif icantly with the spatial outcome measures were entered into separate linear regre ssion analyses using a backward elimination procedure, with the purpose of identifying predictors that si gnificantly account for each spatial outcome variable. Follow-up hierar chical regression analyses were performed using two models. Significant predictors were entered (forced entry) into the first model, unless predictors were theoretically or act ually highly correlated. The second model held these predictors constant and tested the im pact of upper extremit y tremor score to determine whether the change in R2 was significant, and whether the direction of the tremor severity predictor was as hypothesized in relation to t he spatial score in question (positively). Collinearity, resi dual statistics, and leverage va lues also were examined for assumption testing and outliers in these models.
82 CHAPTER 5 RESULTS Specific Aim 1: Group Comparisons on Â“SpatialÂ” and Â“Object/FormÂ” Measures Overview The first aim of this study was to test the hypothesis that individuals with severe upper extremity ET demonstrate s uperior spatial skills relative to those with mild upper extremity ET (see Figure 4-2). Spatial skill s were measured by t he Judgment of Line Orientation and three co mputerized experimental tests: the Visual Search Task (RT, error rate), Mental Rotations Task (RT, error rate), and the Computer-Generated Arena (number of hidden tar gets found, time to reach hidden tar gets, total path length to reach hidden targets, and percent ti me spent in the target qu adrant during the final/probe trial). Group differences were also test ed on visual Â“object/formÂ” tasks (control tasks), wherein no group differences were expecte d to be found. These tasks included the Facial Recognition Test (a neuropsychologic al test) and the Object Attention Task (computerized experimental task). ET par ticipants were not expected to outperform controls on either spatial or object/form visual tasks. Demographic, Cognitive, and Mood Variables Prior to testing these hypotheses, the two ET groups and the controls were compared on demographic variables, cognitive measures (those not in the spatial and object/form batteries), and self-reported mood sco res. As was descr ibed in chapter 4 (Methods), non-significant oneway ANOVAs and chi-square te sts found the mild ET, severe ET, and control groups to be com parable in age, education, gender ratio, handedness, and a full-scale IQ estimate. Addi tionally, in line with expectations, the
83 severe ET group scored higher across tremor seve rity measures relative to the mild ET group. Table 4-2 provides these group characteristics. Controls and the mild and severe ET groups then were compared on measures from the dementia screening te sts, fronto-executive tests, and self-reported anxiety and mood (one-way ANOVAs with Bonferroni-corrected posthoc comparisons, with an alpha criterion of .05). M eans and standard deviations for these measures are provided in Table 5-1. The mild ET, severe ET, and control groups did not differ on anxiety (STAI-State, STAI-Trait), depression (BDI -II), or apathy (Apathy Scale) scores ( p s > .5), but the severe ET group performed relatively poorly on several cognitive measures. Specifically, the severe ET group scored wo rse than controls on the Mini-Mental State Exam, the immediate and delayed recall por tions of the WMS-III Logical Memory subtest, Digit Span, and the Stroop Color-Word in terference condition, whereas the mild ET group performed comparably to controls on these measures. The severe ET group also scored worse relative to the mild ET group on the Boston Naming Test and Category Fluency, although neither ET group differed from the controls on these measures. Scores were within the unim paired range (MMSE scores > 25; other scores > 5th percentile). In sum, the three participant groups were comparable in demographic and mood variables, and the severe ET group scored higher across tremor severity measures relative to the mild ET group. The severe ET group tended to perform statistically worse on dementia screening measures and also demonstrated evidence of mild frontoexecutive functioning deficits (<1 SD di fference when compared with controls),
84 Table 5-1. Means (SD) by group for dementia screening, fronto-executive, and mood variables. Dementia screening measure Control ET mild ET severe Mini-Mental State Exam 29.1 (0.9) 28.7 (1.0) 28.1 (1.5)* WMS-III LM-I (SS) 12.9 (2 .8) 12.8 (2.9) 10.5 (3.3)* WMS-III LM-II (SS) 13.6 (2 .8) 13.4 (2.6) 11.1 (2.6)* Boston Naming Test (T-score) 53.3 (11.9) 58.9 (9.3) 48.6 (11.7)^ Fronto-Executive measures Control ET mild ET severe WMS-III Digit Span Total (SS) 11. 9 (3.6) 11.4 (2.6) 9.3 (2.8)* WMS-III Spatial Span (SS) 12. 2 (2.8) 11.7 (1.6) 10.4 (4.3) Stroop Color-Word (correct ed raw score) 74.1 (11.3) 72.2 (10.9) 65.1 (12.2)* Letter Fluency (FAS) (T-score) 48.1 (7.8) 47.8 (11.2) 42.6 (12.8) Category Fluency (Animals) (T-score) 47.2 (9.4) 52.7 (7.6) 44.0 (10.6)^ Mood/anxiety measure Control ET mild ET severe STAI-State (T-score) 45.1 (7.9) 43.4 (9.1) 46.4 (8.5) STAI-Trait (T-score) 47.1 (8 .8) 45.1 (11.2) 49.5 (12.9) BDI-2 5.4 (4.3) 6.2 (5.5) 6.8 (6.2) Apathy Scale 9.5 (4.0) 10.6 (5.2) 10.8 (6.0) Note: Values are expressed in means (S D). WMS-III = Wechsl er Memory Scale, Third Edition; LM = Logical Memory; SS = sc ale score; TRS = Tremor Rating Scale; STAI = State-Trait Anxiety Scale; BDI-2 = Beck Depression Inventory, 2nd Edition. Severe ET group score < Control group score ( p < .05). ^ Severe ET group score < Mild ET group score ( p < .05).
85 consistent with the lit erature. All individual scores were within the unimpaired range due to the exclusion criter ia that were applied. CRT Results Because some of the experimental m easures are time-dependent, visuomotor reaction time (RT) of the left and right hands was recorded and compared among the ET groups and the controls. This was exam ined using a 3 (Group: control, mild ET, severe ET) x 2 (Hand: left, right) mixed-model ANOVA. The dependent variable was (normal) log10-transformed RT. Results of this ANOVA revealed a significant main effect of Group, F (2,57) = 6.06, p < .01, p 2 = .175. Bonferroni -adjusted pairwise comparisons revealed that RTs were slower for the severe ET group than for the mild ET group ( p < .05, r = .51) and the control group ( p = .01, r = .42). The severe ET group averaged about 0.1 seconds slow er than the other two groups. No difference in RT was found between the mild ET partici pants and control participants ( p > .9), nor were there significant Hand or Hand Group interaction effects ( p s > .3). Means and standard deviations for the RTs are provided in Table 5-2 and depicted in Figure 5-1. Regarding error rates for the left hand and right hand (and the average error rate between the left and right hands), a group di fference was found with a significant Kruskal-Wallis test for the right hand, 2(2) = 146.5, p < .01. This was followed by Mann-Whitney paired comparison s, which found that controls made fewer errors with the right hand than the mild ET group (i.e., hi tting the left button as opposed to the right button), p < .01. Surprisingly, the mild ET group also made more errors with the right hand than did the severe ET group, p < .05. No differences in error rates were found between the controls and the severe ET gr oup (ps > .6) on the right hand. The three groups demonstrated similar error rates wit h the left hand as indicated by the non-
86 Table 5-2. Choice Reaction Time (CRT) task means (SD) by group Measure Hand Control ET mild ET severe CRT RT (seconds) Left Hand 0.62 (0.11) 0.62 (0.08) 0.72 (0.13)* Right Hand 0.61 (0.12) 0.60 (0.08) 0.73 (0.14)* CRT Error Rate Left Hand .013 (.034) .020 (.056) .023 (.044) Right Hand .019 (.053) .053 (.052)^ .008 (.028) Note: "Motor speed" task. RT = Response Ti me (seconds). Error rate range: 0-1. RT: Severe ET group > (Control group, Mild ET group), p< .05 ^ Error rate: Mild ET group > (Control group, Severe ET group), p< .05 Figure 5-1. CRT performance by group, re sponse hand. The severe ET group was slower than the control and mild ET groups. RTs did not vary by hand in any group. For error rates, the mild ET group committed more errors with the right hand than t he other two groups. 0.0 0.2 0.4 0.6 0.8 1.0 ControlET-mildET-severeSeconds 0.0 0.1 0.1 0.2 0.2 ControlET-mildET-severeError Rate Left Right*
87 significant Kruskal-Wallis test in this condition, 2(2) = .78, p = .68. Means and standard deviations for the error rates are provided in Table 5-2 and depicted in Figure 5-1. In sum, the CRT analyses found that the se vere ET group was slower to respond than both the mild ET group and the contro ls, with the mild ET and control groups responding equally fast. The mild ET gr oup made more errors of commission (right hand) than did the severe ET group or the controls. Spatial Tasks (JOLO, MRT, VST, C-G Arena): Group Comparisons JOLO results Raw (and corrected) scores. Data derived from t he JOLO were normally distributed for each subject group and variab le (i.e., non-significant KolmogorovSmirnov tests, p > .05). Thus, a one-way ANOVA wa s performed to compare the raw JOLO scores among the mild ET, severe ET, and control groups. Results of the ANOVA revealed no significant group di fferences in raw JOLO scores, F (2,62) = 1.56, p = .22. This analysis was re-performed on JOLO scores corrected for gender and age (Benton, Sivan, Hamsher, Varney, and Spreen, 1994); however, this result also was not significant, F (2,62) = 2.55, p > .05. Means (SDs) are provided in Table 5-3 and depicted in Figure 5-2. Â“HighÂ” vs. Â“lowÂ” performers. A secondary analysis was performed in order to replicate the approach taken by Springer et al. (2006). They found that a greater-thanexpected proportion of older ET patients fell in the 51st to 100th percentile range on the JOLO (19 of 22 patients). Th is analysis was replicated in the present sample after combining the mild and severe ET groups. The ET participantsÂ’ raw JOLO scores were adjusted for age and education, converted into percentiles, and re-categorized as falling
88 Table 5-3. Judgment of Line Orientat ion (JOLO) means (SD) by group Spatial measure Control ET mild ET severe Test JOLO Raw score 24.0 (4.0) 22.9 (3.8)21.7 (4.6) ns JOLO Corrected score (gender, age) 27.8 (3.9) 26.7 (3.6)25.1 (4.1) ns Note: "Spatial" task. Figure 5-2. Judgment of Line Orientat ion raw scores and ageand gender-corrected scores. No group differences were f ound on either measure. Error bars represent standard deviations. 0 5 10 15 20 25 30 35 Raw ScoreCorrected ScoreJOLO score Control Mild ET Severe ET n s n s
89 (a) above or (b) at/below the 50th percentile. A chi-squared analysis was then performed to compare the act ual number of participants with ET in each group with the expected proportion of ET participants in eac h group (i.e., n/2 in the upper-performing half and n/2 in the lower-performing half). Resu lts were inconsistent with Springer et al. (2006). That is, the ET par ticipants in the present st udy were not, as a group, disproportionately high scorers on the JOLO, 2(1) = 1.20; p = .27, as only 18 of 30 ET patients scored in the 51st to 100th percentile range. In c ontrast, in the study by Springer et al. (2006), 19 of 22 severe ET pati ents fell in this better-performing category. MRT results Response times and error rates on the Mental Rotations Task were compared across four conditions for each group (mild ET severe ET, control). As previously described (see Figure 4-7), the four main c onditions represented hand stimuli rotated at 0, 90-in, 90-out, and 180 degrees. An initial analysis revealed no significant differences in error rates or RTs when left-hand re sponses and right-hand responses were compared; thus, these data were averaged at each angle of rotation. E rror rates and RTs from the MRT are provi ded numerically (Table 5-4) and graphically (Figure 5-3). MRT response times. First, the raw RT data were compared by group and angle of rotation. Non-significant Kolmogorov-Smirnov tests ( p > .05) indicated that each of these variables was normally distributed, for c ontrols as well as both ET groups. A 3x4 mixed-model ANOVA was performed, wit h a between-subjects factor of Group (mild ET, severe ET, control) and a within-subjects factor of Angle (rotated at 0, 90-in, 90-out, and 180 degrees). MauchleyÂ’s test was significant, 2(5) = 53.20, p < .001, indicating that the assumption of sphericity was violated; therefore, Greenhouse-Geisser corrections were applied to the relev ant degrees of freedom.
90 Table 5-4. Mental Rotation Test (MRT) means (SD) by group Measure Angle of Rotation C ontrol ET mild ET severe MRT RT 0-degrees 1.29 (0.55) 1.30 (0.28) 2.08 (1.05) 90-degrees/in 1.47 (0.45) 1.67 (0.39) 2.56 (1.68) 90-degrees/out 1.74 (0.86) 2.31 (0.94) 3.11 (1.84) 180-degrees 2.66 (1.06) 3.65 (1.40) 4.44 (2.26) MRT Error Rate 0-degrees 3.98 (8 .24) 5.86 (7.38) 3.66 (7.03) 90-degrees/in 1.89 (3.31) 4.69 (12.40) 4.97 (5.11) 90-degrees/out 4.55 (13.74) 6.25 (9.68) 6.04 (13.95) 180-degrees 9.28 (12.51) 17.19 (18.33) 7.35 (12.18) Note: "Spatial" task. RT = Response Time (seconds). See Figure 5-3 for significant differences. Figure 5-3. MRT performance by group (contro l, mild ET, severe ET), performance measure (RTs and error rates), and c ondition (angle of rotation). The severe ET group was slower than the mild ET group on the easiest conditions (0 and 90/in), but speed was si gnificantly different in the harder conditions (90/out, 180). Although t he 180 condition was associated with more errors than the other conditions, th is pattern only held for the control and mild ET groups. In this conditi on, theoretically and empirically the most difficult, the seve re ET group outperformed the mild ET group. 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 090/in90/out180 Angle of RotationRT (seconds) 0 0.1 0.2 0.3 0.4 090/in90/out180 Angle of RotationError Rate Controls ET mild ET severe(B) ns < (A) ns ns < < <
91 Results of the ANOVA revealed a significant effect of Group on RT, F (2,58) = 9.40, p < .001, p 2 = .245. In light of significant LeveneÂ’ s tests for the RT variables at all four angles of rotation ( p s < .05), Games-Howell posthoc procedures were used in deconstructing the significant Group effect. Post-hoc co mparisons indicated no significant differences between the two ET groups ( p = .24), though both were significantly slower to respond than the controls ( p s < .05). The effect of Angle on RT also was significant, F (1.83,106.29) = 95.21, p < .001, p 2 = .621, as were all pairwise, post-hoc comparisons between angles of rotation (Bonferroni-adjusted, p s < .001); the fastest times were associated with the 0-degree condition, followed by 90-degrees rotated in, 90-degrees rotated out, and then 180 degrees / upside-down. This pattern was as expected, matching the order of condit ions by increasing di fficulty (i.e., greater effect of biomechanical constraint on visualizing the hand). There was also a significant Angle Group interaction F (3.67,106.29) = 3.28, p < .05, p 2 = .102. Post-hoc comparisons (Bonferroni -adjusted) showed that the severe ET group was slower than the controls and mild ET group in the two easier conditions (0 and 90-in). In the two more difficult c onditions (90-out and 180), the severe ET group was comparable to the mild group (but slower than the controls; p < .01). The mild ET group did not differ in RT from th e control group in any condition ( p s > .05). In sum, the severe ET group was slower t han the mild ET group in the two easier mental rotation conditions (0, 90-in). Howe ver, the two ET groups were comparable in speed during the more difficult conditions ( 90-out, 180). The control group did not differ from the mild ET group in speed.
92 MRT error rates. Group error rates were compared using nonparametric analyses. First, separate FriedmanÂ’s ANOVAs were performed to compare patterns of errors across the four angles of rotation for each participant gr oup. The effect of Angle was significant for the control group, 2(3) = 12.66, p <.01, and the mild ET group, 2(3) = 18.28, p < .001; follow-up Wilcoxon rank-sum tests indicated that the 180-degree angle condition was associated with a signifi cantly higher error rate than the other angles in these two groups ( p s < .05). No differences were apparent among the other anglesÂ’ error rates. While this pattern in e rrors (180 > [0 = 90-in = 90-out]) was found in the controls and participants with mild ET, it was not evident in the severe ET group: a FriedmanÂ’s ANOVA showed that error rates did not vary by angle of rotation among severe ETs, 2(3) = 2.74, p = .43). Mann-Whitney test s compared error rates among the three participant groups in the 180degree condition alone, and found that the severe ET group made fewer errors than the mild ET group ( U = 65.00, p < .05). No other group differences in erro r rates were found in the 180degree condition. In sum, group-wise error rate analyses found that th e 180-degree condition (the theoretically most difficult condition whereby the most mental rotation is required) was associated with relatively more errors than the others fo r both the control and mild ET groups, but not in the severe ET group. Moreover, t he severe ET group outperformed the mild ET group in this condition, with significantly fewer errors. MRT summary. Analysis of these data yielded several interesting findings. Regarding group differences in the speed of re sponses (RTs), the severe ET group was slower than the mild ET group in the tw o fastest conditions (0 and 90 degree inward), but this group difference disappeared in t he two most difficult conditions (90-degree
93 outward and 180 degrees). When error rates were compared across groups and conditions, it was found that the theoretically most difficult condition (180 degrees) was associated with the worst performance in the control group as expected, and this was also the case in the mild ET group. In the severe ET group, however, this pattern was not apparent, with error rates being similar across conditions. When the mild and severe ET groups were compared directly on the 180-degree condition, the severe ET group was found to outperform the mild ET group with significantly fewer errors (despite similar response times in this condition) This pattern was consistent with the hypothesis that spatial skills in severe ET are better than in mild ET, a pattern that emerged only in the most difficult condi tion of the Mental Rotations Test. VST results Next, performances on the Visual Search Task were compared by group (mild ET, severe ET, control) and condition (target tria ls, non-target trials). Â“PerformanceÂ” was based on response times (RT) and error rates. RT and error rate means and standard deviations are transcribed in Table 5-5 and depicted in Figure 5-4, organized by group and condition (i.e., target vs. non-target trials). VST response times. All VST variables were normally distributed after log10 (RT+1) transformations were applied. A 3x2 mixed-model ANOVA was performed with Group as the between-subjects factor and Condition (target trials, non-target trials) as the within-subjects factor. A significant Condition effect was found, F (1,61) = 387.56, p < .001, p 2 = .864, with faster responses being a ssociated with target trials vs. nontarget trials. The effect of Group on RT was significant as well, F (2,61) = 3.49, p < .05, p 2 = .103. Bonferroni-adjust ed pos-hoc analyses demonstrated that the severe ET
94 Table 5-5. Visual Search Test (VST) means (SD) by group. Task Condition Control ET mild ET severe VST RT Target trials 1.82 (0.37) 2.00 (0.43) 2.07 (0.64) Non-target trials2.83 (0.60) 3.01 (0.60) 3.58 (1.08) VST Error Rate Target trials 10. 30 (8.45) 6.41 (6.71) 12.00 (12.68) Non-target trials1.14 (4. 42) 0.78 (1.51) 3.17 (4.86) Note: "Spatial" task. RT = Response Time (seconds). See Figure 5-4 for significant differences. Figure 5-4. VST reaction times (A) and erro r rates (B), by group and condition. The severe ET group was slower and made mo re errors than the controls, but only in the non-target condition. Ac ross groups, targets were recognized faster than non-target tr ials, although targets we re missed (errors) more often than incorrectly recognized as being present. Error bars represent standard deviations. 0.0 1.0 2.0 3.0 4.0 5.0 6.0 TargetNon-TargetSeconds 0.00 0.05 0.10 0.15 0.20 0.25 0.30 TargetNon-TargetError Rates Control ET-mild ET-severe* < >(A) (B) ns ns
95 group was slower than the control group, p < .05, while the mild ET group did not differ from either the controls or the severe ET group. These effects were qualified by a significant Group Condition interaction, F (2,61) = 3.37, p < .05, p 2 = .100, which posthoc comparisons indicated was due to group differences only during non -target trials. Specifically, the severe ET group was signifi cantly slower than the controls in this condition, p < .05. There was also a tendency fo r the severe ET group to be slower than the mild ET group ( p = .09). No difference wa s found between the controls and mild group on the non-target trials, and the th ree groups were comparably fast in the target condition (though a trend was found for t he severe ET group to be slower than the controls, p = .09). No other compar isons were significant or reached trend level. In summary, a comparison of response times by group on the visual search task revealed significantly slower RTs by the severe ET group relative to controls, but only during nontarget trials of t he Visual Search Task. VST error rates. Differences in error rates on the Visual Search Test then were tested between the control, mild ET, and severe ET groups. Non-parametric analyses were performed on these data because they were non-normally distributed despite attempts at normalizing them through tr ansformations. First, three Wilcoxon signedrank tests each compared errors in the targ et and non-target conditions, separately for each of the three participant groups. A Bonf erroni correction was applied, and so all effects are reported at a .0167 level of significance. Fo r all three groups, there were more errors in the target conditi on than the non-target condition ( p s < .006). KruskalWallis tests revealed that, similar to the RT analyses above, error rates differed between the three groups onl y during non-target trials, 2(2) = 6.28, p < .05. Post-hoc
96 Mann-Whitney tests were performed for the nontarget trial conditi on (also Bonferronicorrected such that effects were reported at a .0167 level of signi ficance) and revealed only that the severe ET group made more errors than the controls ( U = 165.00, p = .016, r = -.35). Thus, the only sign ificant group difference in error rates was that the severe ET group made more errors than the control group, and only in the non-target condition. VST summary. Performance comparisons on th e Visual Search Task found no group differences in speed or accuracy in fi nding the target shape on the screen. Group differences did emerge in the absent-target condition; that is, during non-target trials, the severe ET group was relatively slower and less accurate than controls in correctly ruling out the presence of the target shape. The mild ET group performed similarly to the severe ET group as well as controls (RTs error rates), in bot h target and non-target conditions. That the severe ET group performed comparably to the mild ET group neither supported nor c ontradicted the spatial superiority hypothesis in severe ET. The finding that the severe ET group was slower and less accurate than controls during nontarget trials may be a reflection of poorer s patial attention, worse visual discrimination between the similar-looking target and non-target shapes, or simply being more careful in ruling out the presence of a target (e .g., double checking each of the 8 shapes present on the screen). As a whole, parti cipants behaved as expected on the task: participants were more likely to commit errors by missing tar gets that were present than to mistakenly indicate their presence. Regarding speed, participants were quicker to identify a target than to indicate its absence (correct trials).
97 C-G Arena results Motor control (Â“visible targetÂ” trials). Before comparing group performances in spatial memory/navigation on the C-G Aren a task, it was necessary to test for differences in basic motor control of t he thumb-operated control pad (differences in basic motor control on this task would affect interpretation of results involving the spatial memory variables). Two variables were extrac ted from these visible target trials: total path length taken (Â“Arena unitsÂ”) and total latency (seconds) to reach the visible targets. As neither variable was normally distributed, Kruskal-Wallis tests were conducted to determine whether performances varied by group (severe ET, mild ET, controls). Neither test was significant ( p s > .5); thus, the three groups demonstrated equal proficiency in visuomotor control as measur ed by this subtask. Figure 5-5 below depicts these results graphically; means and standard dev iations are provided in Table 5-6. As a follow-up to these analyses, Spearman co rrelations were performed between upper extremity tremor severity and the total lat ency and path length variabl es from the visibletarget trials (among all ET participants). None of these co rrelations were significant ( p s > .1), suggesting that tremor severity did not impact basic use of the controls on the C-G Arena task. These results also were suggesti ve of similar levels of motivation and taskspecific perceptual ability fo r the Â“3-DÂ” virtual environment. Spatial memory (Â“hidden targetÂ” trials). From the eight Â“Hidden TargetÂ” trials portion of the C-G Arena task, four variables were extr acted from the participantsÂ’ performances: (1) total path length across trials (measured in C-G Arena units), (2) total latency across trials (in seconds), (3) number of targets acquired (range: 0-8), and (4) percentage of time in the target quadrant during the ninth, probe trial (after removal of the hidden target). Non-si gnificant Kolmogorov-Smirnov te sts found that all four of
98 Table 5-6. C-G Arena m eans (SD) by group. Condition Measure Control ET mild ET severe Visible Targets Total latency 51. 2 (9.9) 52.4 (12.3) 53.9 (10.2) Total path length 545.2 (58. 8) 540.8 (27.9) 534.4 (23.4) Hidden Targets Total latency 301.8 (117.0) 254.1 (122.4) 297.4 (113.0) Total path length 2,390 (1, 049) 1,924 (782) 2,298 (1,160) Total targets found 6.3 (1.6) 5.3 (2.1) 6.1 (1.2) % time in quadrant* 38.1 (23. 5) 22.6 (25.6) 45.4 (24.9) Note: "Spatial" task. "Hidden Target Trials" refers to the "Spatial Navigation/Memory" portion of C-G Arena, and "Visible Target Trials" refers to the motor control task portion of C-G Arena. Significant differences are described in the text and in Figure 5-6. *Target quadrant, probe trial. Figure 5-5. Motor control on C-G Arena (Â“visible targets trialsÂ”). Analyses revealed no group differences in total path length or total latency to reach the eight visible targets, indicating that the th ree groups could use the controls at a similar level of efficiency. 0 100 200 300 400 500 600 700 Total Path Length (Arena units) Total Latency (seconds) Control ET-mild ET-severe n s n s
99 Figure 5-6 C-G Arena Â“hidden target trialsÂ” performance by group: (A) total path length, (B) total latency, (C) targets found, (D) percentage of time (in the target quadrant during the probe trial). Group differences were found only in Figure D (severe ET > mild ET on this measure of spatial memory for the target location). For (A) and (B ), higher numbers are associated with worse performance, but for (C) and (D ), higher numbers are associated with better performance. 0 1000 2000 3000 4000 ControlET-mildET-severeTotal path length (Arena units) 0 60 120 180 240 300 360 420 480 ControlET-mildET-severeTotal latency (seconds) 0.00 0.20 0.40 0.60 0.80 1.00 ControlET-mildET-severeProportion of probe trial 0 1 2 3 4 5 6 7 8 ControlET-mildET-severeTargets found ns ns ns
100 these variables were normally distributed for the mild ET, severe ET, and control groups. Table 5-6 provides the means and standard deviations for these measures. One-Way ANOVAs were performed on each variable, using Group as the betweensubjects factor. As shown in Figure 5-6, the only signif icant group difference was found for the dependent variable, Â“t ime in the target quadrantÂ” during the probe trial, F (2,62) = 3.59, p < .05, 2 = .076. Bonferroni-corrected pairwis e comparisons indicated that the severe ET group scored better than the mild ET group on this measure (i.e., a higher proportion of time spent in the target quadr ant). No other group comparisons were significant on this measure. In sum, there was some lim ited evidence consistent with the hypothesis that more severe tremor is associated with better spatial functioning; on the C-G Arena task, the severe ET group outper formed the mild ET group on a measure of spatial memory, although no other group differences were indicated from any of the analyses. Object/Form Tasks (FRT, OAT): Group Comparisons FRT results Raw (and corrected) scores. A one-way ANOVA was conducted to test for group differences in raw scores on the shor t-form FRT among the mild ET, severe ET, and control groups. Distributions by gr oup were normal according to KomogorovSmirnov testing ( p s > .05). No significant effect of group was found on the raw FRT scores, F (2,62) = 1.68, p > .05, nor was this analysis significant when it was reperformed on ageand education-corre cted long-form FRT scores, F (2,61) = 1.57, p > .05. Means and standard deviati ons are provided in Table 5-7 and depicted graphically in Figure 5-7.
101 Table 5-7. Facial Recognition Te st (FRT) means (SD) by group. Measure Control ET mild ET severe Raw score (short form) 21.5 (2.4) 22.4 (2.2) 20.9 (2.4) Corrected score (education, age) 45.7 (4.5) 47.4 (4.1) 44.5 (4.9) Note: Object/Form Task. No signific ant group differences on this test as predicted. Figure 5-7. Facial Recognition Test raw scores (short form) and ageand educationcorrected scores (after long-form conversi on). No group differences were found on either measure. Error bars represent standard deviations. 0 10 20 30 40 50 60 Raw ScoresCorrected ScoresFRT Score Control Mild ET Severe ET n s n s
102 OAT results Group differences were tested on four dependent variables extracted from the Object Attention Task. These variables were RTs and error rates when judging two object-based features in two conditions: (1) t he Â“one-objectÂ” condition, in which the two feature judgments were made on a single object and (2) the Â“two-objec tÂ” condition, in which the two feature judgm ents occurred on two different (overlapping) objects but were spatially closer. The raw RT variabl es were not normally distributed for most groups and conditions, but normality for all RT variables was confirmed with Kolmogorov-Smirnov tests ( p > .05) after applying log tr ansformations. Error rates remained non-normal after transformation, and so non-parametric statistics were performed on these data. RT and error ra te values extract ed from the OAT are provided by group in Table 5-8, and di splayed graphically in Figure 5-8. OAT response times. A 3x2 mixed-model ANOVA wa s performed to test group differences in the two conditions. Group (mild ET, severe ET, control) was the betweensubjects factor, and Condition (one object, two objects) was t he within-subjects factor. The main effect of Condition was significant, F (1,60) = 18.43, p < .001, such that RTs were faster for the one-object condition than t hey were for the two-object condition. The effect of Group on RT also was significant, F (2,60) = 5.08, p < .01, p 2 = .235, and Bonferroni-corrected post-hoc analyses indicat ed that the severe ET group was slower than the control group overall, p < .01). No other group differences were significant. The Group Condition interaction also was not significant ( p > .4). In sum, RT analyses on the OAT found that the seve re ET patients were slower than the other groups in making correct feature judgments. The mild ET group did not differ in RT from controls or the ET group with more severe tremor.
103 OAT error rates. The severe ET, mild ET, and control groups also were compared for differences in error rates over the one-object and two-object conditions. Bonferroni-corrected Wilcoxon tests, eval uated at a significance level of .0167, compared error rates in these conditions am ong the three groups. Error rates did not differ significantly between the oneand two-object conditions for any group ( p s > .08). Two Kruskal-Wallis tests found that error rates in both the one-object and two-object conditions did not differ between the three groups ( p s > .2). Thus, error rates did not vary by group or condition on the OAT. OAT results summary. Results of the above analyses showed that error rates did not vary by group or condition, and relati ve to controls, the severe ET group had longer response times regardless of the condition (i.e., whether the perceptual judgments were made on one or tw o objects). The severe ET group did not differ from the mild ET group in RT. Taken together, th ese results are consistent with the initial hypothesis that severe and mild ET would no t differ on this Â“object/formÂ” task, as the two groups, divided by tremor severity, were hypothesized to differ only on spatial tasks. The difference in RT between controls and the severe group was similar in direction and magnitude to that found on t he CRT and therefore appears to reflect merely a group difference in motor RT rather than a subst antial difference in object-based attention. The finding that perceptual judgments in th e two-object condition were slower than those in the one-object condition is consist ent with findings in young adults by Behrman and colleagues (1998), the original authors of this task. Simi lar to the present results, Behrman and colleagues found faster RTs for perceptual judgments wh en the features to be judged occurred on one object. This was es tablished in their study as well as the
104 Table 5-8. Object Attention Test (OAT) means (SD) by group. Task Condition Control ET mild ET severe OAT RT* 1-Object trials 1.05 (0.18) 1.12 (0.21) 1.28 (0.33) 2-Objects trials 1.07 (0.20) 1.14 (0.21) 1.32 (0.35) OAT Error Rate 1-Object trials 1. 12 (2.22) 2.15 (4.03) 2.19 (2.94) 2-Objects trials 0.93 (0.96) 1.76 (4.63) 0.83 (1.43) Note: "Object/Form" test. RT = Response Time (seconds). *The severe ET group was slower than Control group on this task overall ( p< .01), and the 2-Objects condition was associated with slower RTs than the 1-Object condition ( p< .001). No group differences between Severe ET and Mild ET groups, as predicted. Figure 5-8. OAT reaction times and error rate s by group. The severe ET group was, overall, slower than the control group on this test, and the severe ET and mild ET groups were statistically equal in speed. No group differences were found in error rates. n 0.0 0.5 1.0 1.5 One ObjectTwo ObjectsSeconds 0.000 0.020 0.040 0.060 0.080 0.100 One ObjectTwo ObjectsError Rate Control ET mild ET severe ns < ns
105 present one, despite the fact that the two features to be judged always were located substantially further apart in the one-object c ondition relative to the two-object condition (approximately 41% further apart; see Figure 4-10). The present findings support object-based attention in older adults, with or without ET. Summary of Specific Aim Results: Group Comparisons on Spatial vs. Object/Form Tasks This set of analyses compared performances on visual Â“spatialÂ” vs. Â“object/formÂ” measures for groups of individuals with eit her mild, severe, or no ET (controls), with mild vs. severe ET group membership defined by the severity (i.e., amplitude) of upper extremity postural/kinetic trem or. Based on previous findi ngs, it was hypothesized that the severe ET group would out perform the mild ET group on spatially loaded tasks, but not on object/form perception tasks. On the whole, results from these analyse s were partially consistent with this hypothesis. The severe ET group outperform ed the mild ET group on two of the four spatial tasks. Specifically, the severe ET group made fewer errors than the mild ET group on the most difficult condition of t he Mental Rotations Task (180-degrees of rotation), and they also demonstrated better spatial memory on the computer-generated Arena task (i.e., spending more time searching the correct area for the target following the learning trials (1-8). On the other tw o spatial tasks, the mild ET and severe ET groups performed similarly. That is, the two groups identified the presence of a target shape with similar accuracy and speed on the Visual Search Task, and they also demonstrated similar perceptual ability in iden tifying the angular orientations of lines relative to items in a reference array. There was no evidence to suggest that the mild ET group outperformed the severe ET group on any spatial measure. Relative to the
106 controls, neither ET group showed greater accuracy or speed on any of the spatial tests. On the two visual object/form percepti on tasks, the severe and mild ET groups performed comparably, as hypothesized. They demonstrated equal ability in the recognition and discrimination of unfamiliar faces portrayed under normal and unusual lighting conditions. They also perfo rmed comparably (accuracy and RT) on a task requiring efficient perceptual j udgments of overlapping figures As was the case on the spatial tasks, neither of the two ET groups outperformed controls at a statistically significant level on either of these object/form tasks. Analysis of demographic and cognitive measures demonstr ated that the severe ET group and the mild ET group were statistically comparable to each other and controls in age, gender ratio, education, handedness, a full-scale IQ estimate, and mood/anxiety scores. While no member of either ET group was im paired on dementia screening measures (MMSE, Logical Memory subtest of the WMS-III, Boston Naming Test), the severe ET group as a whole scored statis tically worse on the MMSE, Logical Memory immediate and delayed recall, and Boston Naming. Finally, the severe ET group was slower than the mild ET group and contro ls on a measure of simple visuomotor response time using the button box (about 0.1 seconds on average). As a whole, these Â“impairmentsÂ” in the severe ET group were relatively small, but they nevertheless reduce the likelihood t hat the evidence for better spatial performances of the severe ET group vs. the mild ET group on tw o of four spatial tests was due to group differences in some other factor.
107 Specific Aim 2: Controlling for Other Variables in the Spatial Superiority in Severe ET Effect Using Hierarchical Regressions Overview The second aim of this study was to rule out other potential factors underlying the spatial superiority effect in severe ET (m ental rotations, spatial memory/navigation). Analyses in the first specific aim already established that the mild ET and severe ET groups were statistically similar acro ss demographic and mood variables, and the severe ET group scored worse on several cognitive measures (albeit mildly so) and motor speed on the Choice Reaction Time ta sk. Aside from upper extremity tremor severity, then, no other causal factor for the superior spatia l effect in the severe ET group is immediately obvious. The paradoxi cal relationship between higher tremor severity and better spatial skills has not been described previously in the literature, however; thus, the burden of proof is greater. For this reason, the next set of analyses was designed to test more rigorously whether upper extremit y tremor severity significantly and positively predicts spatial functioning in ET above and beyond other predictor variables such as sociodemographic characteristi cs, fronto-executive measures or mood measures. To achieve this end, two hierarchical regr ession analyses were planned, one for each of the two spatial outcome measures whereby the spatial superiority effect in severe ET was demonstrated: (1) t he Mental Rotations Task e rror rate in the 180 degrees condition and (2) the Arena proportion of ti me spent searching the target quadrant during the probe trial. The c ontrols were not included in these analyses, and both the mild and severe ET groups were combined into one sample.
108 Analyses To begin, the MRT and Arena outcome va riables were correlated with upper extremity tremor severity and demographic, cognitive (fronto-executive), and mood variables, in order to determine basic associ ations with spatial functioning. The zeroorder Spearman correlation coefficients are presented in Table 5-9 and characterize relationships between each spatial outcome variable and the potential predictor variables. The set of potential predictor variables included upper extremity tremor severity, demographic variables (gender [r ecoded: 1 = male, 2 = female], age, education, full-scale IQ estimate ), fronto-executive measures (Digit Span total, Spatial Span total, Stroop Color-Word[CW], Le tter Fluency, Category Fluency) and selfreported mood measures (STAI-State, STAI -Trait, BDI-2, and Apathy Scale). As shown in Table 5-9, MRT error rate (180-degrees) was negatively correlated with tremor severity (i.e., fewer MRT errors / higher accuracy was associated with more severe tremor as predicted) It was also positively correlated (i.e., associated with worse performance) with female gender, age, and category fluency ( p s < .05). A negative correlation with Stroop CW (inter ference) was borderline significant ( p = .05). Arena % time was positively correlated with tremor se verity (i.e., better spatial memory / navigation associated with more severe tremor), and negat ively correlated (i.e., associated with better spatial memory) with fe male gender, age, and STAI (state and trait) scores. Finally, corre lations were examined between tremor severity and these (non-spatial) variables that co rrelated significantly with either of the two spatial scores. Tremor severity was only significantly associated (inversely) with category fluency ( r = -.35, p > .05). Tremor severity tended to correlate with male gender ( p > .07), but this did not reach significance.
109 Table 5-9. Correlations with Â“Spatia l SuperiorityÂ” Variables in ET Domain Predictor Correlation with MRT 180 Errors1Correlation with Arena % Time2 Disease Severity Upper Extr emity Tremor -.32* .45** Demographic Gender (1=male, 2=female) .41* -.45** Age .33* -.43** Education .17 .00 FSIQ estimate .08 .04 Fronto-Executive Digit Span total .03 -.02 Spatial Span total -.19 -.01 Stroop CW (interference) -.31^ -.10 Letter Fluency .10 -.03 Category Fluency .29* -.14 Mood/Anxiety STAI-State .13 -.40* STAI-Trait .05 -.33* BDI-2 -.01 -.06 Apathy Scale .11 -.05 1 MRT-180 Errors = Mental Rotations Test, 180-degrees Condition (Error Rate) 2 Arena % Time = Percent of time s pent on probe trial in target quadrant ** p< .01, p< .05, ^p=.05
110 After these analyses were performed, the significant correlates for each of the two spatial measures were entered into s eparate exploratory linear regression analyses using a backward elimination procedure ( one analysis using MRT error rate at 180 degrees as the outcome variabl e, and the other using Arena per cent of probe time in the target quadrant as the outcome vari able). These analyses determined which of these predictors for each spatial variable re mained significantly predictive of spatial functioning when considered together. Follow-up hierarchical linear regression analyses tested the independent c ontribution of upper extremity tremor severity to spatial cognition (MRT or Arena perfo rmance), above and bey ond the predictor variables identified by the ex ploratory regression analyses. Below, the analyses for the MRT outcome variable are presented first, followed by those fo r the Arena outcome variable. MRT (180-degrees condition) regressions Regressing upper extremity tremor se verity, gender, age, category fluency, and Stroop CW scores on MRT error rate with th e backwards elimination procedure found the final significant model, R2 = 0.240, F (2,27) = 3.96, p < .05, with two variables as the only significant predictors of MRT error rate at 180 degrees. These were upper extremity tremor severity = -.37, t (27) = -2.10, p < .05, and Stroop Color-Word = -.37, t (27) = -2.11, p < .05. Subsequently, these two pr edictors were regressed upon MRT performance in a hierarchical regression analysis yielding two models. Results are provided in Table 510. In the first model, the effect of Stroop CW upon MRT errors alone was described. Significance for this m odel reached trend level, R2 = .11, F (1,26) = 3.11, p = .09, with Stroop CW = -.33. In the second model, in wh ich MRT errors were regressed upon
111 both Stroop CW and upper extrem ity tremor severity, 13% additional variance was accounted for, attributable entirely to upper extremity tremor severity, R2-change = .134, F -change(1,25) = 4.40, p < .05. The overall model was significant, R2 = 0.240, F (2,27) = 3.96, p < .05, as were the coefficients for each predictor variable, Stroop CW = -.37, upper extremity tremor = -.37, p s < .05. These results i ndicate that more severe upper extremity tremor significantly predicted fewer errors on the most difficult condition of the mental rotations task, above and bey ond the enhancing effect of cognitive inhibition (as measured by Str oop CW measure) on these kinds of mental operations. Tolerance and VIF values were examined fo r evidence of collinearity, for each of these predictor variables. Both sets of va lues were found to be acceptable, with VIF values = 1.01, and tolerance values greater than .2 (i.e., both .99). None of the residuals in this analysis fell outside a 3-SD range, and all leverage values were acceptable at less than .5 (max value = .33). For these r easons, the model was considered robust and not influenc ed unduly by outlying data points. C-G Arena percent time (probe) regressions Upper extremity tremor severity, gender (male = 1, female = 2), age, and STAIState were regressed upon Ar ena % Quadrant Time using the backwards elimination procedure to determine which variables remai ned predictive of spatial memory when considered together. STAI-Tra it was not included in this analysis because it correlated highly with STAI-Trait scores ( r = .77, p < .01) and was considered to be less theoretically relevant to Ar ena performance (i.e., STAI-State is a measure more closely related to anxiety during task performance, as it measures anxiety in the moment of filling out the questionnaire). This analysis mathematically eliminated gender as a predictor to form a significant model, R2 = 0.482, F (3,27) = 7.45, p < .01, with the
112 remaining predictors being significant: age, = -.56, t (27) = -3.45, p < .01; STAI-State, = -.56, t (27) = -3.47, p < .01; and upper extremity tremor severity, = .39, t (27) = 2.66, p < .05. Age, STAI-State, and upper extremity trem or were then used as predictors in a hierarchical regression with the Arena % ti me being the outcome variable. Age and state anxiety were entered in an initial model and tremor severity in the second to test its independent contribution to the spatial score Results of this analysis are provided in Table 5-10. The first model was significant and accounted for 33% of the variance in spatial memory on the Arena task, R2 = .330, F (2,27) = 6.16, p < .01. Both predictors (age and situational anxiety) were significant in this model ( p s < .01). The second model was also significant, R2 = .482, F (3,27) = 7.45, p = .001, and tremor severity was found to make a significant independent contri bution to spatial memory on this Arena task (15% of the variance), above and beyond age and STAI anxiety/arousal, R2change = .152, F -change(1,2247) = 7.06, p < .05. All three coefficients were significant, with Age = -.56, STAI-State = -.56, UE Tremor = .39; p s < .05. Thus, greater upper extremity tremor predicted significantly better spatial memory performance on the Arena task, above and beyond age (older age being associated with worse spatial memory on Arena) and anxiety/arousal (more si tuational anxiety bein g related to worse spatial memory on Arena) predictors. Collinearity and residual diagnostics were completed for this model as well. VIF val ues were less than 1.21, and tolerance values were above .82. No residuals fell outside 3 SD, and all leverage values were .34 or below (most values < .2). T hese values suggest that the mo del is robust and relatively uninfluenced by outliers.
113 Table 5-10. Summary of Hierarchical Regression Analys es for Variables Predicting Spatial Measures in ET. Model 1 Model 2 Spatial Task Outcome Variable Predictor Variable B SE B Beta B SE B Beta MRT 180-degree error rate Constant 0.30 0.10 0.42 0.098 Stroop Color-Word -0.006 0.003 -0.33 -0.006 0.003 -0.37* Tremor Severitya -0.01 0.006 -0.368* R2 0.11 0.24 F for change in R2 3.11 4.40 C-G Arena % time in quadrant Constant 1.88 0.46 1.8 0.42 Age -0.02 0.005 -0.53 -0.02 0.005 -0.557** STAI-State -0.02 0.005 -0.53 -0.02 0.005 -0.559** Tremor Severitya 0.024 0.009 .392* R2 0.33 0.483 F for change in R2 6.16 7.06 p < .05, ** p < .01 a Based on average TRS scores for left and ri ght upper extremity tremor severity
114 Summary of Results for the Regression Models In sum, these analyses indicated that gr eater severity of tremor in the upper extremities was found to significantly predi ct spatial functioning (mental rotations and spatial memory) in ET, above and beyond other predictors of perform ance. Accuracy of mental rotation in the most difficult condition (rotat ed at 180 degrees) was also positively associated with cognitive inhibition (i.e., the error ra te was negatively associated with inhibition). Better spatial memory/navigation was also positively associated with less situati onal anxiety and younger age.
115 CHAPTER 6 DISCUSSION Study Overview and Results Summary Overview The first specific aim of this study was to test the counterintu itive but theory-driven hypothesis that severe ET is associated with better spatial cognition than milder ET. A sample of non-demented indivi duals with an ET diagnosis was recruited and split into two groups, a Â“mild ETÂ” group with relatively minimal upper extremity intention tremor, and a Â“severe ETÂ” group with more severe upper extremity intention tremor. A control group of healthy, demographically matched subjects without tremor also was recruited. All participants were tested on a series of visuospatial tasks, with each task assessing predominantly spatial judgments or feature-discrimination / identification, both in the visual domain. This distinction follows Un gerleider and MishkinÂ’s (1982) two streams of visual information processing model, which descr ibes a dorsal stream in the brain that processes spatial information, and a ventral stream that processes form information for purposes of object discrimination / identification. In the pres ent study, the Â“spatialÂ” tasks measured the following skills: (1) accuracy of judgment for various linesÂ’ orientations (JOLO), (2) speed and accuracy for searching a field of similar-looking shapes for a target shape (VST), (3) speed and accuracy fo r the mental rotati on of hand stimuli (MRT), and (4) speed and accuracy for learni ng the spatial location of an unmarked target in a virtual room (C-G Arena). The Â“object/formÂ” tasks measured the following skills: (1) accuracy of matching and discrim ination of unfamiliar faces (FRT) and (2) speed and accuracy of form-feat ure judgments for more abstr act figures (OAT). The object/form tasks were included along with the spatial tasks primarily to test that
116 differences in visual cognition in mild vs. severe ET are present only when considering spatial abilities (but not ot her types of visual cognition, such as form/feature-based object identification/discrimination). Support for Superior Spatia l Skills in Severe ET Comparing the performances of the health y control, severe ET, and mild ET groups on the visual cognition tasks yiel ded results that were consistent with hypotheses. First, the severe ET group out performed the mild ET group on two of the four spatial tasks. More specifically, t he severe ET group made fewer errors than the mild ET group on the hardest condition of the mental rotations te st (180 degrees / upside-down), whereas the two groups were co mparable in speed in this condition. The severe ET group also demonstrated better s patial memory than the mild ET group on the C-G Arena task, based on search behavior during the final (probe) trial, which occurred after the initial learning trials. Of note, the severe and mild ET groups and the controls were statistically equal in age, gender, education, a full scale IQ estimate, situational and generaliz ed anxiety, depressive symptom s, and apathy. The three groups also demonstrated similar ability to us e the control pad to navigate as intended on the C-G Arena, and the severe ET group demonstrated the slowest basic reaction time on the CRT. On no task or condition di d the severe ET group perform worse than the mild ET group, and none of the mild ET patients performed bet ter than the severe ET patients on any spatial task. Taken together, these findi ngs did not fall contrary to study hypotheses and yielded only supportive evidence. The second specific aim of this study used regression analysis to determine more specifically whether the bette r spatial performances in the severe ET group (mental rotations, spatial memory) was related to a fact or or factors other t han tremor severity.
117 Again, consistent with hypotheses, tremor severity was found to significantly and positively predict spatial functioning, even beyond the contribution of other significant task-specific predictors (the positive effect of cognitive inhibition). Tremor severity also remained positively predictive of spatial memory, even after accounting for and ruling out the significant, negative effects of age and st ate anxiety on spatial memory. In sum, this series of analyses provided further evidenc e that greater disease severity (i.e., armrelated intention tremor) in ET is, paradoxically, predi ctive of better spatial abilities (i.e., mental rotation of hands and s patial navigation/memory). The Â“controlÂ” tasks categorized as Â“object /formÂ” tasks of visual cognition (Facial Recognition Test, Object Att ention Task) were administered to demonstrate that tremor severity in ET was associated with spatia l skills rather than other types of visual cognition. Indeed, as hypothesized, the severe and mild ET groups performed comparably on these two tests. That is, they were co mparable in accuracy for recognition and discrimination of unfamiliar fa ces, and they were similarly accurate and fast making in feature-based per ceptual judgments of overl apping figures. Neither of the two ET groups outperfo rmed controls on these visual object/form tasks. Some Unexpected Null Findings While the presence and direction of gr oup differences on two spatial tests was found as hypothesized (i.e., the severe ET group being better than the mild ET group on measures of the mental rotations te st and C-G Arena), and the absence of group differences on the two object/form tests were as hypothesized (i.e., the severe ET group performing comparably to the m ild ET group in facial recognition and abstract objectfeature discrimination), expec ted differences on two other spatial measures were not observed by group (i.e., JOLO and VST). On the JOLO, The mean raw score and raw
118 corrected score for the severe ET group was co mparable to the scores of the mild ET and control groups. The fact that the se vere ET group did not demonstrate better performance on this test was contrary to hypotheses and at odds with findings from a previous study. Springer et al. (2006) had found that 19 of 22 pre-surgical ET patients performed above the 50th per centile on this particular test (significantly greater than expected proportion of n/2). In the present study, however, the proportion of patients scoring above the 50th percentil e was not significantly great er than the expected value (18 of 30 total ET patients w ho were tested in the study). On the VST, the severe ET group did not outperform the other two groups in finding a target shape among distractors; rat her, they performed equally, statistically speaking, to the mild ET group and controls The severe ET groupÂ’s performance was comparable to the mild ET gr oupÂ’s performance on the condition of the task whereby no target was present. This condition was in cluded in the task primarily to ensure the participants were attending to the stimuli and completing the task per instructions (i.e., preventing a strategy of simp ly pressing the Â“targetÂ” button as soon as the stimuli were visible rather than finding the target). While the severe ET and mild ET groups had similar performances on this condition of t he VST, i.e., ruling out the presence of the target, the severe ET group was slower and less accurate than the controls. Summary Taken together, this study demonstrat ed that ET patients with severe upper extremity tremor possess similar or better spat ial cognitive skills relative to individuals with mild ET, when demographic factors are comparable and the presence of dementia is ruled out. Better performance was found in the severe ET group on a task of mental manipulation (rotation) of hand representations and another task of spatial
119 navigation/memory. It was not found on ot her spatial tasks including one measuring perceptual judgment/discriminat ion of angles, or another invo lving locating a predefined target shape in a busy field. Moreover, be tter spatial performance in the severe ET group was found only on select measures wit hin the mental rotations and spatial navigation tasks. It was not found pervasive ly across conditions on these tests, and when found, the differences were relatively modest in magnitude. On the measures wherein it was observed, however, it was f ound to be significantly predictive of spatial performance even after controlling for the infl uence of other significant predictors of performance, such as demographic or fronto-executive abilitie s. Furthermore, it should be noted that this spatial superiority in t he severe ET group was established despite several Â“cognitive disadvantagesÂ” in that subgroup, such as poorer memory and executive functioning. No source for the spatial advantage could be found other than tremor severity. Although the severe ET group demonstrated evidence for poorer cognitive skills on several non -visuospatial measures, there are seve ral other factors that might have contributed to the unexpected lack of group differences as hypothesized (severe ET > mild ET) on the Visual Search Task and Judgm ent of Line Orientat ion task. These are discussed below. Potential Factors Accounting for Unexpected Null Findings Visual Search Task Factors A spatial advantage in the severe ET group was not observed on the Visual Search Task. The only group differences that were found included the severe ET group being slower and also making more errors than the control group; this occurred only when the target was not present. The mild ET group, in contrast, performed similarly
120 than controls in this condition. While not overly striking, this pattern of ET group findings relative to controls represents some indirect evidence that the severe ET group may have been slightly worse than the mild ET group in ruling out the presence of a target, although a direct comparison between the two ET groups yielded no differences (on either condition). Even st ill, further discussion of this pa ttern of results is warranted. One possibility for the greater number of non-target errors in the severe ET group (relative to controls) might relate to a higher incidence of simple, executive-motor mistakes, responding with the uni ntended hand. The overall error rate in this condition was quite low, about 5% of trials, which amount s to two trials over 40 in the non-target condition. If this error rate was artificially high, howev er, due to unintentional motor responses due to severe postural tremor, then truly high spatial abilities in the severe ET group might have been masked (as well as for the Mental Rota tion Task). Indeed, the use of motor-based visuospatial tasks pr esents a problem for interpretation when used in the assessment of trem or disorders such as ET or ParkinsonÂ’s disease. The error rate data derived from the Choice Re action Time task, however, do not converge with the idea that executive-motor errors ar e higher in the severe ET group. The CRT data show comparable executive-motor erro r rates on this task when the severe ET group and control groups were compared (about 1.5% of trials). Among the three groups, it was the mild ET group that made more errors on the CRT (right hand), although this error rate was low as well (~3.7% ). Thus, it is possible that on the nontarget condition of the VST, the group differenc e in error rate was merely by chance. The longer response times in the non-tar get condition for the severe ET group (again, relative to the controls) might have reflected greater Â“carefulnessÂ” in ruling out
121 the presence of the target. Subjectively, the experimenter observed some subjects spontaneously reporting that they were doubl e-checking the screens when the target was not found on an initial scan. The CRT da ta are consistent with the idea that the severe ET group tended to be more careful in this way: while the severe ET group was motorically slower than the mild ET group on this task, it made significantly fewer errors on the CRT as mentioned, with a very low error rate overall (averaging one mistake found for every three severe ET participants w ho completed the CRT). This explanation likely applies only to the non-target trial findi ngs, however. The severe ET groupÂ’s error rates and RTs were the same as those for both the controls and the mild ET group in the (main) target condition, and Â“carefulnessÂ” would be reflected by longer response times taken in exchange for lower error rates. A third possibility might relate to the fa ct that performance on this hard-feature serial exploratory visual s earch task more heavily requ ires efficient object/form recognition in addition to spatial attention. The form/shape of each stimulus in the visual field on this task must be compared with the target shape to determine a Â“matchÂ” or Â“mismatchÂ”, and notably, the targets and tw o types of distractors had the same Â“LÂ” shape but different orientations. On non-ta rget trials, the computational load of discriminating between non-target stimuli and the internal r epresentation of the target shape reaches a maximum, as all eight non-target stimuli in this condition must be examined to rule out the presence of t he target. Thus, any advantage in spatial processing that the severe ET group might po ssess would potentially be Â“dilutedÂ” by the non-spatial (i.e., object/form perceptual) elem ents of the VST, with a bias toward more dilution on non-target vs. target trials. Indeed, the authors of this task noted that this
122 task, unlike other feature search tasks, in volves non-covert visual searching (i.e., without fixed gaze) in a serial, item-by-it em manner (Ellison et al., 2004). A linebisection or landmark task might have been more appropriate in retrospect, in that such tasks may be more purely Â“spatialÂ”; this mi ght have yielded group differences in the expected direction. Finally, a spatial advantage was not observ ed on the VST in the severe ET group, potentially because of a problem with Â“domain specificityÂ”. That is, although the VST was chosen for its reliance on right superior te mporal gyrus functioning (Ellison et al., 2004; Gharabaghi et al., 2006), an area that Daniels et al (2006) found to be more cortically dense in severe ET vs. milder ET, it is possible that the relative increase in this area reflected improvement in some other skill related to severe tremor. As pointed out by Daniels et al. (2006), functional neuroima ging studies have shown that this area likely is centrally involved in sensorimotor coordination, the temporal control of movement sequences, goal-direc ted preparatory activity, and control of polyrhythmic movements (Bengtsson, Ehrsson, Forssberg, & Ullen, 2004; Harrington, Rao, Haaland, Bobholz, Mayer, Binder, et al., 2000; Oullier, Bardy, Stoffregen, Bootsma, & 2004; Toni, Shah, Fink, Thoenissen, Passingham, & Zill es, 2002; Ullen, Forssberg, & Ehrsson, 2003). The adjacently located pos terior area of the superior temporal sulcus (pSTS) also is increasingly being appreciated as a cent er for interpreting biological motion, such as articulated movements or intentional actions (as reviewed by Eysenck & Keane, 2005, and Kolb & Whishaw, 2003). Any or all of these skills might be more consistently practiced in this arm-related intentional tr emor disorder via a compensation mechanism (as opposed to visual search for static, relatively meaningless shapes on the VST).
123 Judgment of Line Orientation Test Factors It is unclear as to why better perform ance in the severe ET group was not observed on the JOLO. Springer and coll eaguesÂ’ (2006) study also examined JOLO performance in a severe, older sample of presurgical ET patients. Unlike the present studyÂ’s results, high scores were observed on this test in a disp roportionate portion of the sample (19 of 22 patients scoring rela tively highly compared to ageand gendercorrected norms). The two study samples were comparable in that they were almost identical in age (70 years) and education (14 years), showed evidence of mild frontoexecutive deficits, and similar screening methods were used (i.e., eliminating those with risk factors for generalized cognitive impairment, such as a history of brain injury or failed dementia screening). It is difficult to make direct comparis ons between the present study sample and that tested by Springer et al. (2006), howev er, as other important information was not available in that earlier retrospective database study. Tremor severity was not measured quantitatively by Sprin ger et al. (2006), a factor t hat this study suggests is important for predicting spatial functioning. Instead, Â“severe ETÂ” in the earlier studyÂ’s sample was only assumed, because the patient s were being evaluated for the surgical treatment of presumably impai ring tremor. The region of the body affected by ET was not characterized in the earlier study sample whereas the present study characterized disease severity based on upper extremity tremor Finally, the medication status of the pre-DBS patient sample tested in the earlie r study was not documented. Surgery is typically a treatment option that is reserv ed for medication-refrac tory cases, and a relatively unmedicated sample might have been abl e to perform at a level closer to their true abilities. Superior spatial functioning would have been more evident in such a
124 sample, but these are simply conjectures. In any case, anti-tr emor medications, especially in higher doses, can have general ized negative effects (i.e., across tests). Medication status and its potential negative influence on spatial functioning is considered in more detail in the next section. This is followed by a discussion of other Â“generalÂ” factors that could systematic ally account for poorer-than-expected performance on the JOLO, VST, and other spatial tests. Anti-Tremor Medications Two of the most commonly administered pharmacologic agents for the treatment of ET include primidone and propranolol, both of which in higher doses can be associated with mild-to-moder ate drowsiness, dizziness, and confusion (Zesiewicz et al., 2008). It is possible that the severe ET group would have outperformed the mild ET group more pervasively across spatial measur es, if cognitive side effects of these and other tremor-reducing medications (e.g., al prazolam, topiramate) were a substantial factor in this study. Indeed, deleterious e ffects of tremor medication on cognition might have accounted for the slightly worse fronto-executive defici ts observed in the severe ET group relative to the other groups. The proportion of the severe ET group that reported taking medications to reduce tremor (11:4, or 73.3%) was not statistically greater than the propor tion of the mild ET group reporting the use of antitremor medications (9:7, or 56.4%). The proportion was numerically greater in the severe ET group, however, and it is possible that the severe group was taking heavier doses of these medications (or ta king more types simultaneously) to treat more severe tremor. If so, cognitive im pairment (or even mild cognitive disturbances) due to these medications would be more likely in the severe ET group. In turn, performance accuracy and/or speed would be reduced on spatial
125 measures in this study, potentially masking spatial superiority effects on measures whereby this was not actually observed. Unfortunately, there is no direct method of comparison between the effects of different medications with different doses in different individuals. Participants were asked during screening whether they felt their tremorreducing medications impacted their thinking. The majority of participants denied this, while the remainder indicated their uncertainty None definitively indicated this in the affirmative. Ideally, the participants in this study w ould have been unmedica ted so that the effects of tremor medication could be ruled out. They were not asked to discontinue them prior to their partici pation, however, because medi cation washout periods are usually 1-3 weeks in ET studies (e.g., Gir onell, Kulisevsky, Barbanoj, Lopez-Villegas, Hernandez, Pascual-Sedano, 1999; Milanov, 2002). This duration of time off medications likely would have hampered ET re cruitment efforts substantially, especially for severe ET patients. Although the impact of medications on spatial cognition is only inferred (as disproportionately affecting the severe ET group negatively), this group still outperformed the mild ET group on the m ental rotations test and the spatial navigation/memory test. Motor Requirements Several efforts were made to minimize any tremor-related effects on experiments using motor-based responses in this study. Tests that lacked motor-based responses were selected when possible, and for the experimental tasks with motor response requirements (i.e., use of t he button box or the control pad), the input devices were chosen and placed in such a way that ET pa rticipants could minimize the interfering effects of tremor. Control tasks for motor responses were implemented to measure any
126 residual differences. Despite these efforts, the motor-based require ments in the present study potentially represent a met hodological limitation to the inte rpretation of the results. Importantly, these challenges were more lik ely to interfere negativ ely with the spatial performances of the severe ET group, thus biasing results aw ay from the directions of the studyÂ’s hypotheses (i.e., severe ET > m ild ET on spatial tasks). Nevertheless, the overall pattern of results might have been cl earer if voiceor foot-activated response devices were used for the reaction-time based measures (although the use of these devices would not be free from criticism either). Theoretical Assumptions of the Study Functional Organization of the Visual Brain The fact that only two of four Â“spatialÂ” tasks yielded findings directly supporting the study hypotheses might be related to in correct assumptions about the functional organization of the visual brain. The Â“obj ect/formÂ” and Â“spatialÂ” task distinction in the present study followed the theoretical distin ction between the func tions of the ventral (occipito-temporal) and dorsal (occipito-parie tal) streams of information processing (Ungerleider & Mishkin, 1982). This model pr oposes that the ventral stream processes forms, shapes, and/or identity, whereas the dorsal stream processes information about spatial relationships. This Â“what-whereÂ” model was historically influential but not free from criticism in more recent years. From a theoretical st andpoint, Ettlinger (1990) noted that objects and forms (e .g., size, shape, pattern) are themselves defined by spatial relationships and ther efore these two constructs are fundamentally inextricable for object perception. Another functional model of the visual brain (Milner & Goodale, 1992; 2009) attempted to address this concern and other limitations of Ungerleider and MishkinÂ’s
127 (1982) model. Milner and Gooda le noted that Ungerleider and MishkinÂ’s (1982) Â“whereÂ” vs. Â“whatÂ” model of visual processing fails to account for more recent findings relating to dorsal stream processing. Their extensiv e study of a patient with ventral stream damage shows that despite blindness for visual recognition of objects, this patient can Â“unconsciouslyÂ” reach for objects and also shape her hand in the correct way for grasping the object, regardless of the objectÂ’s identity. T hey also observed that in patients with dorsal stream damage, conscious recognition of objects and their features (including spatial properties) are intact, but there is impairment in manipulating the hands in the correct way during reach-to-gra sp tasks. Even in the earlier work upon which Ungerleider and Mishkin based their t heory, Â“spatialÂ” deficits were defined in monkeys who were impaired on reac hing and grasping behaviors. These observations and subsequent, conver gent findings led Milner and Goodale (1992) to spearhead a proposal that the dorsal visual stream (which projects toward somatosensory and motor cortices) may be be tter characterized as a set of systems used for the online, moment-by-moment visual control of action. The ventral visual stream (which projects towa rd inferior temporal cortex) handles the type of visual perception used in deriving t he structure of the environm ent, such as processing and identifying objects from relatively invari ant form-based and spatial properties (Milner & Goodale, 2009). Put simply, these two visual systems are sometimes called the Â“vision for actionÂ” and Â“vision for percept ionÂ” streams of processing. Though not infallible, Milner and GoodaleÂ’s functional framework might be useful for understanding why the severe ET group out performed the mild ET group on the C-G Arena and the Mental Rotations Task, but not JO LO or the Visual Search Task. Using
128 this model to reclassify these tests, the C-G Arena task and Mental Rotations Task are, arguably, tests requiring more action-based vis ual processing. The Visual Search Task and JOLO, on the other hand, ar e relatively visual-per ceptual tasks (as are the previously categorized Â“object/formÂ” ta sks: Object Attention Task and Facial Recognition Test). Specifically, on the C-G Arena task, spat ial memory performance is dependent upon the efficiency of visuomotor control. As they propel themselves forward in the circular arena, participants must use const antly updating visual cues to determine current location and heading (i.e., Â“optic flowÂ”). They calibra te current direction Â“on the flyÂ” with fine-motor adjustment s in order to guide themselves accurately to the desired location. For these reasons, performance on the C-G Arena seems theoretically reliant on the vision for action system, at least fo r efficient egocentric navigation (allocentric mapping is mediated more by entorhinal, hippocampal, and parahippocam pal cortices). On the version of the Mental Rota tion Task, performance is based on the embodiment and mental, visuom otor manipulation of hand-bas ed stimuli. There was strong evidence that mental rotation of these hands wa s egocentric and motor based, as suggested by the increasing RTs for conditi ons requiring more ment al rotation (i.e., 180 > 90-in or 90-out > 0 degrees) and thos e limited by actual biomechanical constraints (i.e., 180 > 90-out > 90-in > 0 degrees ) in controls. Arguably, the simplest condition on this task (i.e., 0-degrees) is merely a perceptual condition whereby no mental rotation is performed, and in this condition, the severe ET group did not outperform the mild ET group. This group di fference was evident only in the condition that required the greatest degree of mental visuomotor manipulation (i.e., 180 degrees).
129 These results support the idea that the rotation conditions of the MRT utilize Â“vision for actionÂ” functions. While it might be argued that no overt action actually occurs during the MRT, Glover (2004) descr ibed evidence demonstrating that visual planning of action as well as visual contro l of action both utilize relatively dorsal areas (i.e., superior parietal lobe and inferior parie tal lobe, respectively). Other evidence suggests that an inferior parietal area, the s uperior marginal gyrus, is crucial for motor imagery (mental rotation of hand drawings), whereas the superior parietal lobes is critical for non-embodied visual imagery (mental rotation of letters) (Pelgrims, Andres, & Olivier, 2009). Regardless of the distinction between overt and covert action, in mental rotation, the vision for action pathway (or at least its functional domain) is more heavily involved during performance on hand-based MRT than ventral (perceptual) areas. In severe forms of ET, then, Milner and GoodaleÂ’s (1992) Â“vision for actionÂ” system arguably is the visual system that under goes the most use and development via compensation for severe arm tremor, relative to milder forms of ET or living without tremor. From a functional perspective, thos e with severe tremor must expend greater effort in the visuomotor guidance or control of the upper extremitie s during their use. For this reason, it is perhaps not surprising that group differences were not found on the JOLO and the (serial-perceptual) Visual Search Test (or the Object Attention Task or the Facial Recognition Task), in light of this more updated functional model of the visual brain. Behavioral Neuroplasticity Driving the Spatial Superiority Effect? The data supporting behavioral neuroplasticity as the mechanism for (visuomotor) spatial superiority in severe ET is inferred because of limitations to the study design. This study gathered no structural neuroimaging data with which to correlate to spatial
130 performances. Thus, the findings presented by Daniels and colleagues (2006) showing greater cortical density in visuospatial areas (severe ET > mild ET) could not be replicated here, and it was only assumed that this pattern app lied to the present studyÂ’s ET sample. Moreover, no longitudinal dat a were gathered in this study, rendering impossible any direct relationship between in creasing tremor severity and better spatial performances within the same individuals. The cross-sectional design was favored in this study for pragmatic reasons; the time-c ourse of substantially increasing tremor often spans greater than a dec ade in ET (though as noted above, the trajectory can vary considerably among individuals). Neve rtheless, causality ca nnot be assumed with this correlational design. Moreover, other potentially in fluential factors might have accounted for the spatial superiority effect in severe ET (above and beyond tremor severity). The regression analyses attempted to identify factors other than tremor severity that might better account for the visuospatial advantage demonstr ated in the severe ET sample. Other potentially causal variables were not measur ed, and for those that were, the regression analyses were limited a relatively small numbe r of predictors (N=3) because of the small sample size in the study. One small possib ility is that the severe ET group happened to possess better spatial skills than the mild ET group even prior to t he onset of tremor. This might be ruled out in future studi es with an analysis of pre-tremor hobbies, vocations, and/or school subjects. These da ta were not gathered with enough detail in this study, though it is unlikely to be the case in actuality, given the randomized nature of recruitment, the equal distributions of gender and educational level across groups, the small likelihood that be tter spatial skills somehow causes more severe tremor, and
131 the converging evidence from Springer et al (2006) and Daniels et al. (2006) as well as the present study that all s uggest the positive relationship between arm tremor severity and better spatial skills. Regardless of the mechanism behind the spatial advantage demonstrated in the severe ET group, the p henomenon itself is given more support with these data (especially in conjunction with the previous literature), which suggests that this represents a promising area for further investigation. Other Directions for Future Research Additional descriptive studies are needed to replicate this behavioral phenomenon in other ET samples and to more compr ehensively describe the parameters under which it operates. There are many outstanding questions. Does the positive relationship between tremor severity and s patial functioning exist on other tasks requiring mental or overt visuomotor manipulat ion or guidance? Is it only observed on tasks related to the affected area of tremor (upper extremities) or can it be established using other targets of embodiment or control (e.g., the m ental rotation of feet, or spatial navigation using foot-operated controls)? Do skills gener alize to non-embodied entities (e.g., mental rotation of three-dimensiona l geometric figures)? One might answer some of these questions by experimentally manipul ating both the rotation strategy (egocentric vs. allocentric) and the stimulus type ( hands. vs. feet vs. geom etric figures). These and other studies would be relati vely theory-driven because it would use the well-supported distinction between Â“vision for actionÂ” vs. Â“vision for perceptionÂ” streams of processing. The work by M ilner and Goodale (for review, see Milner & Goodale, 2009) used several tasks to test these distinct aspec ts of vision using patients with well-defined lesions. These can be classi fied into one or the other category (based on the performance of their patients) and used to test the spatial superiority effect in
132 severe ET. Finally, still other studies might incorporate tasks with more dynamic stimuli and ecological validity. These might ai m to determine whether certain spatial advantages have bearing on activities of daily living, for example. Aside from these additional, theory-driven descriptive studies, future research should attempt to more directly examine ne urobiological changes in spatial areas of the brain along with improvements on cognitive / behavioral measures. This might be achieved by using structural and/or func tional neuroimaging in tandem with cognitive testing, within the context of a longitudinal de sign. A longitudinal study design obviously would be laden with several pr agmatic challenges, including cost and participant dropout, especially over a relatively (but necessarily) long time period between measurements. One advant age, however, is that behavioral changes could be assessed directly within the same indivi duals and possibly yield compelling results, especially if tremor severity increases were found to relate to improvements in subtypes of spatial cognition as well as their underlying neural correlates. The pragmatic difficulties, however, may be too re strictive to conduct such a study. An alternative possibility is to study the relationship between structural brain changes, tremor severity, and spatial cognition in a healthy population. Tremor might be artificially simulated using motoriz ed, weighted mechanism s on the arms, and various tasks could be practiced while this trem or is induced. Â“DosageÂ” (i.e., intensity of disturbance) and duration could be manipulat ed experimentally using this type of paradigm, effectively accelerating the timet able of such a study. This manipulation would also grant more precis e experimental control and allow one to characterize the levels of severity and duration of tremor needed to induce a measurable effect in spatial
133 functioning. The use of healthy subjects woul d allow the experiment er to rule out the effects of medications or neur ological conditions, and they w ould likely be easier to recruit. These and other future studies also would benefit fr om more objective measurement of intentional tremor. EMG, accelerometers, high-s peed video recording, or sensor-based 3-D motion tracking and model ing are several options. This would increase the reliability of this me asure and reduce error-related bias. Aside from the use of heal thy controls, additional st udies might examine the relationship between tremor severity and spatia l functioning in other tremor disorders such as cerebellar stroke or resection, or Pa rkinson disease (PD). It is intriguing that visuospatial deficits are a common compla int in PD, another age-related movement disorder that, unlike ET, is often characte rized by tremor at rest but not during intentional movement (Grow don & Corkin, 1987). Like ET, the characterization of visuospatial functioning in PD probably has been biased toward impairment because of methodological issues including test validity, the use of timed or motor-based measures, or inattention to the cognitive effects of medication status, depr ession, dementia, or comorbid neurological disease (Cruci an & Okun, 2003; Brown & Marsden, 1986; Waterfall & Crowe, 1995). However, highpowered, methodologica lly well-controlled studies testing a wide range of visual f unctioning in PD still have demonstrated impairments across even elementary visual-per ceptual abilities such as contrast sensitivity, visual/spatial attention, and s patial motion and perception (e.g., Uc, Rizzo, Anderson, Quian, Rodnitzky, & Dawson, 2005) The nature of the relationship between spatial functioning and tremor severity in PD (in the resting tremor subtype) is unclear and could be tested directly to clarify the mechanism of the effect observed in the
134 present study with ET. That is, if increasing severity of resting tremor is associated with visuospatial decline in PD, then this could prov ide evidence that adj ustment to intention tremor is necessary for improvement in spatia l skills, consistent with the view presented here that it is the Â“vision for actionÂ” f unctions that undergo behavioral neuroplasticity. Such a study would need to be well contro lled to rule out factors associated with general cognitive impairment. Significance of the Study These findings demonstrated fo r the first time a significant positive contribution of tremor severity for spatial cognition (ment al rotation and spatial memory/navigation), above and beyond other predictors. This yielded behavioral evidence that in conjunction with previous behav ioral and structural neuroimagi ng findings is consistent with a compensation / behavioral neuroplasticity hy pothesis. That is, it is hypothesized that adjustment to living with intentional arm tremor is a ssociated with improvement of certain action-related visuospatial skills via compensation, and that this improvement is mirrored by brain-related neuroplastic chang es. These data imply that cognitive rehabilitation and training visuos patial skills is possible in this and potentially other ageassociated neurological populations.
135 APPENDIX A TREMOR RATING SCALE
136 TRS (continued): reading sample When sunlight strikes raindrops in the air, they act like a prism and form a rainbow. The rainbow is a division of white light into many beautiful colors. These take the shape of a long round arch, with its path high above, and its two ends apparently beyond the horizon. There is, according to legend, a boiling pot of gold at one end. People look, but no one ever finds it. When a man looks for something beyond his reach, his friends say he is looking for the pot of gold at the end of the rainbow.
141 APPENDIX B C-G ARENA PARAMETERS AND OBJECT TEXTURES Visible Targets Trials (Non-Memory Based Motor Control Condition): Parameters Room, Arena Wall Parameters Room Width Room Depth Room Height Arena Radius Arena Wall Height Arena # of Sides 150.0 150.0 45.0 50.0 3.50 60 Start Positions / Orientations, Target Positions Trial Start X,Y* Start Area (Ori entation) Target X,Y* Location** Practice 50,0 South 0 (North) 50,50 Center 1 50,0 South 0 (North) 50,50 Center 2 0,50 West 270 (W) 75,75 Northeast 3 50,100 North 180 (S) 25,25 Southwest 4 100,50 East 90 (E) 25,50 West 5 50,0 South 90 (E) 25,75 Northwest 6 0,50 West 180 (S) 75,25 Southeast 7 50,100 North 270 (W) 75,50 East 8 100,50 East 0 (N) 50,25 South *Coordinates expressed in percentages of t he Arena diameter, with X origin = East, Y origin = South ** Target size is 15x15 and circular in shape User Parameters View Height Field of View Move Quantum Turn Quantum 3.00 90.00 0.50 1.50 Room Schematic (example shown = Trial 6) North 100 West 50 East 0 0 50 100 South Legend Symbol Start/Orientation > Example Path ------End X Target O
142 Visible Targets Trials: Object Textures North Wall East Wall South Wall West Wall Floor Ceiling Arena Wall
143 Hidden Target Trials (Spatial Memory / Navigation Condition): Parameters Room, Arena Wall Parameters Room Width Room Depth Room Height Arena Radius Arena Wall Height Arena # of Sides 150.0 150.0 45.0 50.0 3.50 60 Start Positions / Orientations, Target Positions Trial Start X,Y* Start Area (Orientation) Target X,Y* Location** Practice 50,0 South 0 (North) None None 1 50,0 South 0 (North) 26.5,71.5 Northwest 2 0,50 West 270 (W) 75,75 Northwest 3 50,100 North 180 (S) 25,25 Northwest 4 100,50 East 90 (E) 25,50 Northwest 5 50,0 South 90 (E) 25,75 Northwest 6 0,50 West 180 (S) 75,25 Northwest 7 50,100 North 270 (W) 75,50 Northwest 8 100,50 East 0 (N) 50,25 Northwest Probe 85,85 Northeast 225 (SW) None None *Coordinates expressed in percentages of the Arena diameter, with X origin = East, Y origin = South ** Target size is 15x15 and circular in shape (when visible, after acquisition) User Parameters View Height Field of View Move Quantum Turn Quantum 3.00 90.00 0.50 1.50 Room Schematic (hidden target has fixed location) North 100 West 50 East 0 0 50 100 South
144 Hidden Target Trials: Object Textures North Wall East Wall South Wall West Wall Floor Ceiling Arena Wall
145 LIST OF REFERENCES Aguirre, G. K., Detre, J. A., Alsop, D. C., & DÂ’Esposit o, M. (1996). The parahippocampus subserves t opographical learning in man. Cerebral Cortex 6 823Â–829. Amorim, M., Isableu, B., Jarr aya, M. (2006). Embodied spat ial transformations: Â“body analogyÂ” for the mental rotation of objects. Journal of Experimental Psychology, 135, 327Â–347. Astur, R. S., Taylor, L. B ., Mamelak, A. N., Philpott, L. Sutherland, R. J. (2002). Humans with hippocampus damage display se vere spatial memory impairments in a virtual Morris water task. Behavioural Brain Research, 132, 77Â–84. Ball, K. K., Beard, B. L., Roenker, D. L., Miller, R. L., & Griggs D. S. (1988). Age and visual search: expanding the useful field of view. Journal of the Optical Society of America. A, Optics and Image Science, 5, 2210Â–2219. Ball, K. K., Berch, D. B., Helm ers, K. F., Jobe, J. B., Leveck, M. D., Marsiske, M., et al. (2002). Effects of cognitive training inte rventions with older adults: a randomized controlled trial. Journal of the American M edical Association, 288, 2271Â–2281. Barrash, J., Damasio, H., Adolphs, R. & Tranel, D. (2000). The neuroanatomical correlates of route learning impairment. Neuropsychologia, 38, 820Â–836. Barona, A., Reynolds, C. R., & Chastain R. (1984). A demographically based index of premorbid intelligence for the WAISÂ–R. Journal of Consulting and Clinical Psychology, 52, 885Â–887. Beck, A. T. (1996). Beck depression inventory (2nd ed.) USA: The Psychological Corporation. Bengtsson, S. L., Ehrsson, H. H., Forssberg, H., & Ullen, F. (2004). Dissociating brain regions controlling the te mporal and ordinal struct ure of learned movement sequences. European Journal of Neuroscience, 19, 2591Â–2602. Benito-Leon, J., & Louis, E. D. (2006). Essential tremor : emerging views of a common disorder. Nature Clinical Practice Neurology, 2, 666Â–678; quiz 662p following 691. Benito-Leon, J., Louis, E. D., & Bermejo -Pareja, F. (2006a). Population-based casecontrol study of cognitive f unction in essential tremor. Neurology, 66, 69Â–74. Benito-Leon, J., Louis, E. D., & Bermejo-P areja, F. (2006b). Elder ly-onset essential tremor is associated with dementia. Neurology, 66, 1500Â–1505. Benton, A. L., Hamsher, K., Varn ey, N. R., & Spreen, O. (1983). Judgment of line orientation New York: Oxford University Press.
146 Benton, A. L., Van Allen, M. W., Hams her, K. de S., & Levin, H. S. (1978). Test of Facial Recognition, Form SL: Manual Iowa City, Iowa: Benton Laboratory of Neuropsychology. Benton, A. L., Sivan, A. B., Hamsher, K., Varney, N. R., & Spreen, O. (1994). Contributions to Neuropsychologic al Assessment: A Clinical Manual New York: Oxford University Press. Behrman, M., Zemel, R. S., Mozer, M. C. (1998). Objectbased attention and occlusion: evidence from normal participant s and a computational model. Journal of Experimental Psychology and Human Perceptual Performance, 24, 1011Â–1036. Bethell-Fox, C. E., Shepard, R. N. (1988). Mental rotation: e ffects of stimulus complexity and familiarity. Journal of Experimental Psychology, 14, 12Â–23. Brown, R. G., & Marsden, C. D. (1986). Visu ospatial function in Parkinson's disease. Brain, 109, 987Â–1002. Carroll, J. B. (1993). Ab ilities in the domain of visual perception. Human cognitive abilities: a survey of factor-analytic studies (pp. 304Â–363). New York: Cambridge University Press. Cicerone, K. D., Dahlberg, C., Malec, J. F. Langenbahn, D. M. Felicetti, T., Kneipp, S., et al. (2005). Evidence-based cognitive rehabilitation: updated review of the literature from 1998 through 2002. Archives of Physical Medicine and Rehabilitation, 86, 1681Â–1692. Crucian, G. P. & Okun, M. S. (2003). Vis ual-spatial ability in ParkinsonÂ’s disease. Frontiers in Bioscience, 1, S992Â–S997. Daniels, C., Peller, M., Wolff, S., Alfke, K., Witt, K., Gaser, C., et al (2006). Voxel-based morphometry shows no decreases in cerebel lar gray matter volume in essential tremor. Neurology, 67, 1452Â–1456. Desmurget, M., Bonnetblanc, F., & Duffau, H. (2007). Contrasting acute and slowgrowing lesions: a new door to brain plasticity. Brain, 130, 898Â–914. Deuschl, G., Bain, P., & Brin, M. (1998) Consensus statement of the Movement Disorder Society on Tremor. Ad Hoc Scientific Committee, S13, 2Â–23. Dogu, O., Sevim, S., Camdevir en, H., Sasmaz, T., Bugdayci, R., Aral, M., et al. (2003). Prevalence of essential tremor: door-to-door neurologic exams in Mersin Province, Turkey. Neurology, 61, 1804Â–1806. Draganski, B., Gaswer, C., Busch, V., Schui erer, G., Bogdahn, U ., May, A. (2004). Neuroplasticity: changes in grey matter induced by training. Nature, 427, 311Â–312.
147 Duane, D. D., & Vermilion, K. J. (2002). Co gnitive deficits in patients with essential tremor. Neurology, 58, 1706. Dupuis, M. J., Delwaide, P. J., Boucquey, D., & Gonsette, R. E. (1989). Homolateral disappearance of essential tremor after cerebellar stroke. Movement Disorders, 4, 183Â–187. Edwards, J. D., Wadley, V. G. Myers, R. S., Roenker, D. L. Cissell, G. M., & Ball, K. K. (2002). Transfer of a speed of processing intervention to near and far cognitive functions. Gerontology, 48, 329Â–340. Ekstrom, A. D., Kahana, M. J. Caplan, J. B., Fields, T. A ., Isham, E. A., Newman, E. L., & Fried, I. (2003). Cellular networ ks underlying human spat ial navigation. Nature, 425, 184Â–188. Elble, R. J. (2000). Essential tr emor frequency decreases with time. Neurology, 55, 1547Â–1551. Ellison, A., Schindler, I., Patt ison, L. L., & Milner, A. D. (2004). An expl oration of the role of the superior temporal gyrus in visual search and spatial perception using TMS. Brain, 127, 2307Â–2315. Epstein, R., DeYoe, E. A., Press, D. Z., Rosen, A. C., & Kanwisher, N. (2001). Topographical learning mechani sms in parahippocampal cortex. Neuropsychology, 18, 481Â–508. Epstein, R. & Kanwisher, N. (1998). A cortical represen tation of the local visual environment. Nature, 392, 598Â–601. Ettlinger, G. (1990). Â“Object visionÂ” and Â“spatial visionÂ”: the neuropsychological evidence for the distinction. Cortex, 26, 319Â–331. Eysenck, M. W. & Keane, M. T. (2005). Cognitive psychology: A studentÂ’s handbook (5th Ed.) New York: Psychology Press. Fahn, S., Tolosa, E., & Marin, C. (1993). Clinical rating scale for tremor. In J. Jankovic & Tolosa, E. (Eds.), Parkinson's Disease and Movement Disorders (pp. 225Â–234). Baltimore: Williams & Wilkins. Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). "Mini-mental state". A practical method for grading the cognitive stat e of patients for the clinician. Journal of Psychiatric Research, 12, 189Â–198. Funk, M., Brugger, P. (2008). Mental rotation of congenitally absent hands. Journal of the International Neuropsychological Society, 14, 81Â–89.
148 Gasparini, M., Bonifati, V., Fabrizio, E., F abbrini, G., Brusa, L., Lenzi, G. L., et al. (2001). Frontal lobe dysfunction in ess ential tremor: a preliminary study. Journal of Neurology, 248, 399Â–402. Gharabaghi, A., Fruhmann Berger, M., Tatagiba, M., & Karnath, H. O., (2006). The role of the right superior temporal gyrus in vis ual search Â– insights from intraoperative electrical stimulation. Neuropsychologia, 44, 2578Â–2581. Gironell, A., Kulisevsky, J., Barbanoj, M. Lpez-Villegas, D., He rnndez, G., PascualSedano, B. (1999). A random ized placebo-controlled comparative trial of gabapentin and propranolol in essential tremor. Archives of Neurology, 56, 475Â– 80. Glover, S. (2004). Separate visual representations in the plann ing and control of action. Behavioral and Brain Sciences, 27, 3Â–78. Golden, C. J. (1978). St roop color and word test. A manual for clinical & experimental users Wood Dale, IL: Stoelting. Goldman-Rakic, P. (1987). Circuitry of pr imate prefrontal cort ex and regulation of behavior by representational memory. In V. B. Mountcastle (Ed.), Handbook of physiology (Vol. 5, pp. 382Â–392) New York: Raven Press. Green, C. S., & Bavelier, D. (2003). Action video game m odifies visual selective attention. Nature, 423, 534Â–537. Gron, G., Wunderlich, A. P., Spitzer, M., Tomczak, R., & Riepe, M. W. (2000). Brain activation during human navigation: Gender-d ifferent neural networks as substrate of performance. Nature Neuroscience, 3, 404Â–408. Growdon, J. H., & Corkin, S. (1987). Cogniti ve impairments in Parkinson's disease. Advances in Neurology, 45, 383Â–392. Grutzendler, J., Kasthuri, N., & Gan, W. B. (2002). Long-term dendritic spine stability in the adult cortex. Nature, 420, 812Â–816. Harrington, D. L., Rao, S. M., Haaland, K. Y., Bobholz, J. A., Mayer, A. R., Binder, J. R., & Cox, R. W. (2000) Specialized neural systems unde rlying representations of sequential movements. Journal of Cognitive Neuroscience, 12, 56Â–77. Heaton, R. K., Miller, S. W., Taylor, M. J., Grant, I. (2004). Revised comprehensive norms for an expanded Halstead-Reitan battery: demographically adjusted neuropsychological norms for African American and Caucasian adults. Lutz, Florida: Psychological A ssessment Resources, Inc. Higginson, C. I., Wheelock, V. L., Levine, D., King, D. S., P appas, C. T., Sigvardt, K. A. (2008). Cognitive deficits in essential tr emor consistent with frontosubcortical dysfunction. Journal of Clinical and Experimental Neuropsychology, 15, 1Â–6.
149 Jacobs, W. J., Laurance, H. E., & Thomas, K. G. F. (1997). Plac e Learning in Virtual Space I: Acquisition, overshadowing, and transfer. Learning and Motivation, 28, 521Â–542. Jacobs, W. J., Thomas, K. G. F., Laurance, H. E., & Nadel L. (1998). Place Learning in Virtual Space II: Topographical Relations as One Dimension of St imulus Control. Learning and Motivation, 29, 288Â–308. Jordan, K., Heinze, H.-J., Lutz, K., Kanow ski, M., & Jancke, L. (2001). Cortical activations during the mental rotation of objects. Neuroimage, 13, 143Â–152. Kaplan, E. F., Goodglass, H., & Weintraub, S. (1978). The Boston naming test Philadelphia: Lea & Febiger. Kempermann, G., Kuhn, H. G., & Gage, F. H. (1997). More hippocampal neurons in adult mice living in an enriched environment. Nature, 386, 493Â–495. Kim, J. S., Song, I. U., Shim, Y. S., Park, J. W., Yoo, J. Y ., Kim, Y. I., et al., (2009). Cognitive impairment in ess ential tremor without dementia. Journal of Clinical Neurology, 5, 81Â–84. Kleim, J. A. & Jones, T. A. (2008). Principles of experience-dependent neural plasticity: implications for rehabilit ation after brain damage. Journal of Speech, Language, and Hearing Research, 51, S225Â–S239. Klein, D. A., Steinberg, M., Galik, E., Steele, C., Sheppard, J. M., Warren, A., et al., (1999). Wandering behaviour in comm unity-residing persons with dementia. International Journal of Geriatric Psychiatry, 14, 272Â–279. Kolb, B. & Whishaw, I. Q. (2003). Fundamentals of human neuropsychology (5th Ed.). New York: Freeman-Worth. Koller, W. C., Busenbark K., & Mi ner, K. (1994). The relationship of essential tremor to other movement disorders: report on 678 patients. Essent ial Tremor Study Group. Annals of Neurology, 35, 717Â–723. Kramer, A. F., Bherer, L., Colcombe, S. J., Dong, W. & Greenough, W. T. (2004). Environmental influences on cognitive and brain plasticity during aging. Journal of Gerontology A, Biological Sciences and Medical Sciences, 59, M940Â–957. Kumru, H., Begeman, M., Tolosa, E., Valls-Sol e J. (2007). Dual task interference in psychogenic tremor. Movement Disorders, 22, 22077Â–22082. Lacritz, L. H., Dewey, R., Jr., Giller, C., & Cullum, C. M. (2002). Cognitive functioning in individuals with "benign essential tremor. Journal of the International Neuropsychological Society, 8, 125Â–129.
150 Lewis, C. B. (2002). Aging: The health-care challenge (4th ed.) Philadelphia: F. A. Davis Company. Lohman, D. F., & Nichols, P. D. (1990). Training spatial abilitie s: effects of practice on rotation and synthesis tasks. Learning and Individual Differences, 2, 67Â–93. Lombardi, W. J., Woolston, D. J., Roberts, J. W., & Gross, R. E. (2001). Cognitive deficits in patients wit h essential tremor. Neurology, 57, 785Â–790. Louis, E. D. (2006). Essential tremor. Clinical Geriatric Medicine, 22, 843Â–857, vii. Louis, E. D., Barnes, L., Albert, S. M., Cote L., Schneier, F. R., Pu llman, S. L., et al. (2001). Correlates of functional di sability in essential tremor. Movement Disorders, 16, 914Â–920. Louis, E. D., Faust, P. L., Vons attel., J. P., Honig, L. S., Ra jput, A., Robinson, C. A., et al. (2007). Neuropathological changes in essential tremor: 33 cases compared with 21 controls. Brain, 130, 3297Â–3307. Louis, E. D., Ford, B., & Frucht, S. (2003) Factors associated with increased risk of head tremor in essential tremor: a comm unity-based study in northern Manhattan. Movement Disorders, 18, 432Â–436. Louis, E. D., Marder, K., Cote L., Wilder, D., Tang, M. X ., Lantigua, R., et al. (1996). Prevalence of a history of shaking in per sons 65 years of age or older: diagnostic and functional correlates. Movement Disorders, 11, 63Â–69. Lyons, K. E., & Pahwa, R. (2004). Deep brain stimulation and essential tremor. Journal of Clinical Neurophysiology, 21, 2Â–5. Lyons, K. E., & Pahwa, R. (2005). Handbook of essential tr emor and other tremor disorders New York: Taylor & Francis. Maguire, E. A., Gadian, D. G. Johnsrude, I. S., Good, C. D., Ashburner, J., Frackowiak, R. S. J., Frith, C. D. (2000). Navigation -related structural change in the hippocampi of taxi drivers. Proceedings of the National Academy of Sciences, 97, 4398Â–4403. Milanov, I. (2002). Clinical an d electromyographic assessment of essential tremor treatment. Parkinsonism and Related Disorders, 8, 343Â–348. Milner, A. D., & Goodale, M. A. (1992). Separate visual pathways for perception and action. Trends in Neuroscience, 15, 20Â–25. Milner, A. D., & Good ale, M. A. (2009). The visual brain in action Oxford University Press: USA. Moghal, S., Rajput, A. H., DÂ’A rcy, C., & Rajput, R. (1994) Prevalence of movement disorders in elderly community residents. Neuroepidemiology, 13, 175Â–178.
151 Morris, R.G. (1981). Spatial localization does not require the presence of local cues. Learning and Motivation, 12, 239Â–260. Morris, R.G. (1984). Developments of a water-maze procedure for studying spatial learning in the rat. Journal of Neuroscience Methods, 11, 47Â–60. Morris, R.G., Garrud, P., Rawlin s, J. N., & OÂ’Keefe, J. (1 982). Place navigation impaired in rats with hippocampal lesions. Nature, 297, 681Â–683. OÂ’Craven, K. M., Downing, P. E., & Kanwishe r, N. (1999). fMRI evidence for objects as the units of attentional selection. Nature, 401, 584Â–587. OÂ’Keefe, J. & Dostrovsky, J. (1971). The hi ppocampus as a spatial map: preliminary evidence from unit activity in the freely moving rat. Brain Research, 34, 171Â–175. Okun, M. S., Rodriguez, R. L., Mi kos, A., Miller, K., Kellison, I., Kirsch-Darrow, L., et al. (2007). Deep brain stimulation and t he role of the neuropsychologist. The Clinical Neuropsychologist, 21, 162Â–189. Oullier, O., Bardy, B. G., St offregen, T. A., & Bootsma, R. J. (2004). Task-specific stabilization of postural coor dination during stance on a beam. Motor Control, 8, 174Â–187. Pahwa, A., Lyons, K. E., & Pahwa, R. (2005). Medical treatment of essential tremor. In K. E. Lyons & R. Pahwa (Eds.). Handbook of essential tremor and other tremor disorders (pp. 389). New York: Taylor and Francis. Pahwa, R., Lyons, K., & Koller W. C. (2000). Surgical treat ment of essential tremor. Neurology, 54, S39Â–44. Pelgrims, B., Andres, M., & Olivier, E. (2009). Double dissociation between motor and visual imagery in the po sterior parietal cortex. Cerebral Cortex, 19, 2298Â–2307. Putzke, J. D., Whaley, N. R., Baba, Y., Wszolek, Z. K., & Ui tti, R. J. (2006). Essential tremor: predictors of disease progression in a clinical cohort. Journal of Neurology, Neurosurgery, and Psychiatry, 77, 1235Â–1237. Ratcliff, R. (1993). Methods for dea ling with reaction time outliers. Psychological Bulletin, 114, 510Â–532. Rautakorpi, I., Takala, J., Mart tila, R. J., Sievers, K., & Ri nne, U. K. (1982). Essential tremor in a Finni sh population. Acta Neurologica Scandinavica, 66, 58Â–67. Sahin, H. A., Terzi, M., Ucak, S., Yapici, O., Basoglu, T., & On ar, M. (2006). Frontal functions in young patients with essentia l tremor: a case comparison study. Journal of Neuropsychiatry and Clinical Neuroscience, 18, 64Â–72.
152 Schneier, F. R., Barnes, L. F., Albert, S. M. Louis, E. D. (2001). Characteristics of social phobia among persons with essential tremor. Journal of clinical psychiatry, 62, 367Â–372. Shepard, R. N., & Metzler, J. (1971). Mental rotation of three-dimensional objects. Science, 17, 701Â–703. Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., & Jacobs, G. A. (1983). Manual for the state-trait anxiety inventory Palo Alto, CA: Consulting Psychologists Press. Spreen, O., & Benton, A. L. (1977). Neurosensory Center for Comprehensive Examination for Aphasia (NCCEA). Vi ctoria: University of Victoria Neuropsychology Laboratory. Springer, U., Chang, Y., Graf-Radf ord, J., Jacobson, C., Cruci an, G. P., Okun, M. S., et al. (2006). Neuropsychological Profile of Essential Trem or. Paper presented at the Annual Meeting of the Am erican Psychological Association, New Orleans. Thomas, K. G. F., Hsu, M., Laurance, H. E. Nadel, L., & Jacobs, W. J. (2001). Place learning in virtual space III: investigati on of spatial navigati on training procedures and their application to fMRI and clinical neuropsychology. Behavior Research Methods, Instruments, and Computers, 33, 21Â–37. Toni, I., Shah, N. J., Fink, G. R., Thoenissen, D., Passingham, R. E., & Zilles, K. (2002). Multiple movement representations in the human brain: an ev ent related fMRI study. Journal of Cognitive Neuroscience, 14, 769Â–784. Trachtenberg, J. T., Chen, B. E., Knolt, G. W., Feng, G., Sanes J. R., Welker, E., et al. (2002). Long-term in vivo imaging of exper ience-dependent synaptic plasticity in adult cortex. Nature, 420, 788Â–794. Treisman, A., & Gelade, G. (1980). A feature integration theory of attention. Cognitive Psychology, 12, 97Â–136. Troster, A. I., Woods, S. P., Fi elds, J. A., Lyons, K. E., Pahwa, R., Higginson, C. I., et al. (2002). Neuropsychological defic its in essential tremor: an expression of cerebellothalamo-cortical pathophysiology? European Journal of Neurology, 9, 143Â–151. Uc, E.Y., Rizzo, M., Anderson, S. W., Qian, S., Rodnitzky, R. L., & Dawnson, J. D. (2005). Visual dysfunction in Parkinson disease without dementia. Neurology, 65, 1907Â–1913. Ullen, F., Forssberg, H., Ehrss on, H. H. (2003). Neural networks for the coordination of the hands in time. Journal of Neurophysiology, 89, 1126Â–1135.
153 Ungerleider, L. G., Courtney S. M., & Haxby, J. V. (1998). A neural system for human visual working memory. Proceedings of the National Ac ademy of Sciences USA, 95, 883Â–890. Ungerleider, L. G., & Haxby, J. V. (1994). 'What' and 'w here' in the human brain. Current Opinion in Neurobiology, 4, 157Â–165. Ungerleider, L. G., & Mishkin, M. (1982). Ingle, D. J., Goodale, M. A., Mansfield, R. J. W., eds. Analysis of Visual Behavior MIT Press. Vingerhoets, G., Lange, F. P. Vandemaele, P., Deblaere, K., & Achten, E. (2002). Motor Imagery in Mental Rotation: an fMRI Study. Neuroimage, 17, 1623Â–1633. Waterfall, M. L., & Crowe, S. F. (1995). Meta-analytic co mparison of the components of visual cognition in Parkinson's disease. Journal of Clinical and Experimental Neuropsychology, 17, 759Â–772. Wechsler, D. (1997a). Wechsler Adult Intelligence Scale (3rd Ed.). San Antonio, TX: The Psychological Corporation. Wechsler, D. (1997b). Wechsler Memory Scales (3rd Ed.). San Antonio, TX: The Psychological Corporation. Wills, A. J., Jenkins, I. H., Thompson, P. D., Findl ey, L. J., & Brooks, D. J. (1994). Red nuclear and cerebellar but no olivary acti vation associated with essential tremor: a positron emission tomographic study. Annals of Neurology, 36, 636Â–642. Yantis, S. & Serences, J. T. (2003). Cortical mechanism s of space-based and objectbased attentional control. Current Opinion in Neurobiology, 13, 187Â–193. Young, A. W., De Haan, E. H., Bauer, R. M. (2008). Face perceptio n: a very special issue. Journal of Neuropsychology, 2, 1Â–14. Zacks, J.M. (2008). Neuroimaging studies of mental rotation: a meta-analysis and review. Journal of Cognitive Neuroscience, 20, 1Â–19. Zesiewicz, T. A., Elble, R., Loui s, E. D., Hauser, R. A., Sulliv an, K. L., Dewey, Jr., R. B., et al. (2005). Practice param eter: Therapies for essent ial tremor: Report of the Quality Standards Subcommittee of t he American Academy of Neurology. Neurology, 64, 2008Â–2020.
BIOGRAPHICAL SKETCH Utaka Springer was born in Menomonie, Wisc onsin. He received his Bachelor of Science degree in biology from Harvard Univer sity. He is a doctoral candidate in the Department of Clinical and Health Psychology at the University of Florida, where he is specializing in neuropsychology. He is cu rrently a neuropsychology intern at the Veterans Affairs Northern California Heal th Care System (VA NCHCS) in Martinez, California (September 2009Â–August 2010). He has accepted a research and clinical neuropsychology postdoctoral position with VA NCHCS and the Universi ty of California at Berkeley. This dissertation is an extension of the i deas originating from a project presented at the American Psychologic al Association annual confer ence in New Orleans, 2006, under the mentorship of Dr. Dawn Bowers. This project was funded in part by the American Psychological Foundation through the 2007 Benton-Meier Neuropsychology Award.