The nature of movement and action errors produced by brain-injured patients

MISSING IMAGE

Material Information

Title:
The nature of movement and action errors produced by brain-injured patients
Physical Description:
xiv, 134 leaves : ill. ; 29 cm.
Language:
English
Creator:
Macauley, Beth Lynn Martin, 1965-
Publication Date:

Subjects

Subjects / Keywords:
Communication Processes and Disorders thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Communication Processes and Disorders -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 125-131).
Statement of Responsibility:
by Beth L. Macauley.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 029549529
oclc - 41372660
System ID:
AA00020704:00001

Full Text













THE NATURE OF
MOVEMENT AND ACTION ERRORS
PRODUCED BY BRAIN-INJURED PATIENTS












BY

BETH L. MACAULEY


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1998




























Copyright 1998

by

Beth L. Macauley




























For Shawn,
My gift from God













ACKNOWLEDGEMENTS

I would like to begin by extending a very special thank you to Leslie J. Gonzalez-

Rothi, Ph.D., for her unwavering support and encouragement as I pursued the doctoral

degree. I can remember her coming into my "office" during my VA traineeship in 1989

and saying, "You should get a Ph.D. Think about it." Since that time, her mentorship

has been invaluable as she imparted her extensive knowledge in the realm ofneurogenic

communication disorders, her expertise in patient care and being an advocate for her

patients, as well as her personal morals and standards on being a well-rounded individual

and dedicated wife and mother. I am honored to have her as not only a mentor but also a

friend.

Another special thank you goes to Kenneth M. Heilman, M.D., not only for his

support as I served as a research assistant under his grants, but also for imparting his

expertise with great enthusiasm to those working under his supervision. Dr. Heilman's

encouragement and support allowed undergraduate students, graduate students, neurology

residents, and neurology, psychology, and speech pathology faculty to work together in a

smooth, caring, "family" atmosphere which is rare and unique in the realm of higher

education. My skills and knowledge and ability to work in collaborative teams, as well

as my respect for other disciplines, are the direct result of Dr. Heilman's influence and

guidance.

Other special thank yous go to Bruce Crosson, Ph.D., Linda J. Lombardino,

Ph.D., and Ira Fischler, Ph.D., for their invaluable comments, suggestions, and









encouragement given as part of my doctoral committee. These three elite faculty

members have been superb teachers as well as supporters during my graduate career.

I can not say enough about my faculty colleagues in the Department of Speech

and Hearing Sciences at Washington State University. My chairman, Gail D. Chermak,

Ph.D., is a strong advocate for junior faculty and was gracious in allowing me a reprieve

of some faculty duties in order to concentrate and complete this dissertation. Her

unwavering encouragement and support has been a blessing to me as I made the

transition from graduate student to faculty. The other members of the department, Jeanne

Johnson, Ph.D., Chuck Madison, Ph.D., Tony Seikel, Ph.D., Leslie Power, M.S., and

Linda Vogel, M.S., have also given of themselves and served as mentors and friends

during my initial years in their department. Their encouragement, support, and faith that

I would finish the dissertation were never far from my mind. I can honestly say that

without their "reality checks," continuous encouragement, and willingness to ease my

load by increasing theirs, I would not have been able to finish.

More thank yous go to my graduate students, Holly Wiseman, Stacy Wendel,

Maria McClelland, and Erin Beneteau, at University Programs in Communication

Disorders (UPCD), a cooperative graduate program in speech-language pathology and

audiology between Washington State University and Eastern Washington University.

These students willingly tolerated the training protocols and were genuinely excited to be

taking part in research opportunities and endeavors.

I also wish to thank Robert Short, Ph.D., of the Washington Institute for Mental

Illness Research and Training (WIMIRT) for his assistance with the statistics.








A huge thank you goes to my husband often years, Shawn P. Macauley, Ph.D.,

for his unwavering encouragement, prods in the rear to keep moving along, and for never

doubting that I could get it done. His ability to encourage me in the midst of dealing with

an acquired disability that affected his vestibular system and resulted in physical, visual,

and cognitive impairments, has been a blessing above and beyond expectations. His

unconditional love, support, and assistance allowed me to focus on writing without

feeling too guilty about not keeping up with family responsibilities. Thank you also to

my girls, seven year old Erin and three year old Emily, and my boy, three month old

Evan, for their unconditional love as well as their acceptance of "Mommy's computer

time," take-out twice a week, and piles of dirty laundry. Erin and Emily were always

ready with a hug and a kiss whenever they were needed (and even when they were not

needed) while Evan's smiles melted my heart and put everything in perspective.

I also wish to thank my parents, John and Barbara Martin of Bradenton, Florida

for their unconditional love, unwavering support and encouragement throughout my

college career. They never doubted my abilities or decisions and were always looking for

ways to help and show their support. I also wish to thank my in-laws, Curtis and Karen

Winters of North Bangor, New York, for always being a phone call away to impart

guidance, encouragement, and support through the dissertation years.

Last, but certainly not least, I wish to thank my Lord and Savior, Jesus Christ, for

His personal involvement in my life. I am nothing without Christ and my work is a

testimony to the grace that He has given. I honestly believe that without Christ walking

by my side and carrying me when necessary, I would not have finished this research

endeavor, especially with the life-changing trials that have occurred within the past year.









He started me on this journey and He has never failed to be there throughout. A special

thank you to my dear friend, Priscilla Welboumrn, who never let me forget what was most

important in life and encouraged me to keep my feet firmly planted on the Rock while

reaching for the stars. Her unconditional friendship and unwavering support during the

final writing stages of this dissertation has been a blessed and treasured gift from the

Lord.













TABLE OF CONTENTS

Page

ACKNOW LEDGEM ENTS..................................................................................................... iv

LIST OF TABLES......................................................................................................... x

LIST OF FIGURES ..................................................................................................... xii

ABSTRACT............................................................................................................. xiii

CHAPTER

ONE REVIEW OF THE LITERATURE AND STATEMENT OF
THE PROBLEM .................................................................................................. 1

Introduction......................................................................................................... 1
Review of the Literature..................................................................................... 3
Historical Perspectives and Theoretical
M odels of Apraxia........................................................................... 3
Cognitive Neuropsychological Model
Of Lim b Praxis and Apraxia.................................................... 12
Right Hemisphere Contributions to Praxis.............................................. 19
The Ecology of Apraxia........................................................................... 21
Significance to Clinicians and Patients........................................... 21
Apraxia in Natural Contexts............................................................ 23
The Relationship Between Apraxia and Pragmatic Action..................... 38
Predictions of the Proposed Relationship Between
M ovements and Actions.................................................................. 49
Classification of Action Errors................................................................. 53
Statement of the Problem ...................................................................................... 56

TW O M ETHODS.......................................................................................................... 63

Subjects............................................................................................................... 63
M materials and M ethods ....................................................................................... 67
Evaluation of Independent Variables....................................................... 68
Disorders of Action Planning and Organization................................ 68
Disorders of Learned, Skilled M ovement.......................................... 69
Disorders of Tool-Object-Function Knowledge................................ 72
viii








Disorders of Supervisory Attention/Working Memory..................... 73
Disorders of Language....................................................................... 74
Neglect................................................................................................ 74
Evaluation of Dependent Variable........................................................... 75
Rater Training ..................................................................................................... 79
Statistical Analyses............................................................................................. 80

THREE RESULTS .......................................................................................................... 82

Experim ental Results......................................................................................... 82
Research Questions............................................................................................. 85
Research Question #1............................................................................... 85
Research Question #2a ............................................................................ 87
Research Question #2b............................................................................. 93
Summ ary of Findings........................................................................................... 98

FOUR DISCUSSION ..................................................................................................... 101

Research Questions............................................................................................. 103
M ethodological Issues and Lim stations of the Study......................................... 111
Implications for Future Research....................................................................... 113
Clinical Implications.................................... 116

APPENDICES

A GESTURE TO COMMAND SUBTEST RANDOMIZED
FORM SA&B ....................................................................................................... 119

B TOOL-OBJECT M ATCHING TEST.................................................................. 123

C STANDARDIZED SCORES FOR TRAILS A & B........................................... 124

REFERENCES.................................................................................................................. 125

BIOGRAPHICAL SKETCH ............................................................................................ 132











ix














LIST OF TABLES


Table Pa2e

2-1 LBD Subject Identification...............................................................................65

2-2 RBD Subject Identification ...................................................... ..................... 66

2-3 Control Subject Identification......................................................................... 67

3-1 Descriptive Statistics for Each Group on Measures of
Independent Variables........................................................ ............. .......84

3-2 Overall Kruskall-Wallace ANOVA for Errors
Produced During the Meal ................................................ .............................. 85

3-3 Comparison of LBD and RBD Groups with Normal Controls ............................ 86

3-4 Pearson Product Moment Correlation Report ......................... ............................. 88

3-5 Average Number of Errors Produced by All Subjects when
Divided into Two Groups According to Cut-off Score
For Each Test................................................................................................... 89

3-6 Average Number of Errors by Specific Error Type Produced
By LBD, RBD, and Control Groups............................................................... 90

3-7 Results of Multiple Stepwise Regression Analysis Between
Independent Variables and Total Errors Produced During
the M eal........................................................................................... ................. 94

3-8 Results of Multiple Stepwise Regression Analysis Between
Independent Variables and Tool Errors Produced During
the M eal............................................................................................................ 94

3-9 Results of Multiple Stepwise Regression Analysis Between
Independent Variables and Non-Tool Errors Produced During
the M eal ........................................................................................................... 95








3-10 Results of Multiple Stepwise Regression Analysis Between
Independent Variables and Misuse Errors Produced During
the M eal.............................................................. ............................................ 95

3-11 Results of Multiple Stepwise Regression Analysis Between
Independent Variables and Misselection Errors Produced During
the M eal................................................................................ ...................... 96

3-12 Results of Multiple Stepwise Regression Analysis Between
Independent Variables and Movement Errors Produced During
the M eal ................................................................. .. .......... ......................... 96

3-13 Results of Multiple Stepwise Regression Analysis Between
Independent Variables and INT Errors Produced During
the M eal................................................................. .......................................... 97

3-14 Results of Multiple Stepwise Regression Analysis Between
Independent Variables and Sequence Errors Produced During
the M eal........................................................................................................... 97

3-15 Results of Multiple Stepwise Regression Analysis Between
Independent Variables and Timing Errors Produced During
the M eal ......................................................................... .................................98

3-16 Results of Multiple Stepwise Regression Analysis Between
Independent Variables and Quantity Errors Produced During
the M eal.................................................................... ....................... ............. 98














LIST OF FIGURES


Figure Page

1-1 Liepmann's M odel of Praxis.................................................................... 6

1-2 Geschwind's Model of Praxis................................................................... 9

1-3 Heilman and Rothi's Model of Praxis..................................................... 11

1-4 Cognitive Neuropsychological Model of Apraxia................................... 13

1-5 Delineation of Gesture to Command Pathway........................................ 14

1-6 Delineation of Gesture to Visually Presented Object Pathway ............... 18

1-7 Schematic Representation of Misuse and Mis-Selection
A action Errors............................................................................................ 28

1-8 Schematic Interpretation of the Unified Hypothesis................................ 36

1-9 Relationship Between Actions and Movements...................................... 40

1-10 Breakdown of Low Level Schemas......................................................... 41

1-11 Breakdown of High Level Schemas ........................................................ 44

1-12 Addition of Perceptual Input to Low Level Schemas.............................. 46

1-13 Addition of Semantic System to High Level Schemas ............................ 47

1-14 Proposed Relationship Between Apraxia and Pragmatic Action ............48

2-1 The Tower of London .............................................................................. 70













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE NATURE OF
MOVEMENT AND ACTION ERRORS
PRODUCED BY BRAIN-INJURED PATIENTS

By

Beth L. Macauley

August 1998

Chair: Leslie J. Gonzalez-Rothi, Ph.D.
Major Department: Communication Processes and Disorders

Some patients with brain damage produce action errors during activities of daily

living. Investigators in recent research reports have examined the planning, organization,

and production of these action errors through different theoretical models with different

inherent assumptions and predictions, using different research hypotheses and evaluation

criteria, and including subjects with heterogeneous etiologies of brain damage. The

current study describes a proposed relationship between apraxia and pragmatic action that

serves as a framework for the examination of the nature of movement and action errors

produced by brain-injured patients. The hypothesized relationship between apraxia and

pragmatic action can be divided into two parts--high level schemas associated with

executive function, supervisory attention/working memory, and the conceptual

knowledge of the relationship between tools and actions, and low level schemas

associated with the praxis system which mediates learned skilled movements. Forty

xiii








subjects participated in this study--twenty left brain damaged patients (LBD), ten right

brain damaged patients (RBD), and ten non-brain damaged control subjects. All subjects

underwent an evaluation that measured high and/or low schema deficits and were

videotaped while engaged in an activity of daily living, eating a meal. The videotapes

were scored for presence of action and movement errors. Results indicated that both

brain-injured groups significantly differed from the control group in total number of

errors produced. The LBD group produced more tool errors than the RBD group while

both the LBD and the RBD group produced significantly more non-tool errors than the

control group but did not differ from each other. Additionally, production of specific

error types during the meal correlated highly with a measure of high level schema deficits

and a measure of low level schema deficits. Therefore, both high and low level schema

deficits contribute to production of action errors. Additionally, the results appear to

indicate that brain damage to either hemisphere can result in production of action errors

in the natural environment.













CHAPTER ONE
REVIEW OF THE LITERATURE AND
STATEMENT OF THE PROBLEM

Introduction

Purposeful movements, defined as intentional "changes in place, position, or

posture" (Webster, 1981, p.747), that have direct purpose or aim, and goal-directed

actions, defined as "acts of will accomplished over a period of time in stages"

(Webster, 1981, p. 12) are learned throughout life as people interact with their

environment. The loss of the ability to produce purposeful, skilled movements as the

result of brain damage has been termed apraxia (Rothi & Heilman, 1997). Liepmann

(1920) defined apraxia as an impairment in the production of learned (or skilled)

movements not caused by weakness, paralysis, incoordination, or sensory loss

(Liepmann, 1920). Apraxia is manifested in a person's inability to "move the moveable

parts of the body in a purposeful manner even though motility is preserved" (Liepmann,

1900/1977, p.161).

A deficit in the ability to produce goal-directed actions has been called frontal

apraxia by Mayer, Reed, Schwartz, Montgomery, and Palmer (1990), Schwartz, Reed,

Montgomery, Palmer, and Mayer (1991), Schwartz, Mayer, DeSalme, and Montgomery

(1993), Schwartz, Montgomery, DeSalme, Ochipa, Coslett, and Mayer (1995), and

Schwartz and Buxbaum (1997) and a deficit in managerial knowledge by Sirigu,

Cohen, Duhamel, Pillon, Dubois, and Agid (1995), Sirigu, Zalla, Pillon, Grafminan,

Agid, and Dubois (1995, 1996), and Sirigu, Zalla, Pillon, Grafminan, Dubois, and Agid








(1996). Specifically, both frontal apraxia and deficits in managerial knowledge were

defined as impairments in the planning and sequential hierarchical organization of

actions required to obtain a desired goal (Schwartz & Buxbaum, 1997; Sirigu et aL.,

1995).

Mayer et al. (1990), Schwartz et al. (1991, 1993, 1995), and Schwartz and

Buxbaum (1997) reported that brain damaged patients with frontal apraxia produce

errors of action while performing goal-directed activities of daily living such as

preparing coffee and brushing teeth. These learned, goal-directed actions were

subsequently labeled "pragmatic action" by Schwartz and Buxbaum (1997). Ochipa,

Rothi, and Heilman (1989) reported a brain damaged patient with ideational/conceptual

apraxia who made errors of action while eating a meal and brushing teeth, both

"pragmatic actions as defined by Schwartz and Buxbaum (1997). Foundas, Macauley,

Raymer, Maher, Rothi, and Heilman (1995) reported that brain damaged patients with

ideomotor apraxia made action errors while performing the pragmatic action of eating a

meal. Each of the above studies documented action errors in goal-directed activities of

daily living. However, each of the above researchers approached the study of action

and action errors through different theoretical models with different predictions and

assumptions. Mayer et al. (1990), Schwartz et al. (1991, 1993, 1995), and Schwartz

and Buxbaum (1997) approached the study of pragmatic action through the Unified

Hypothesis, a theoretical model based on the activation-trigger-schema (ATS)

framework proposed by Norman and Shallice (1986). The ATS framework for

production of action was derived from the study of normal subjects (Norman &

Shallice, 1986). Ochipa et al. (1989) and Foundas et al. (1995) approached the study of








pragmatic action through the Cognitive Neuropsychological Model of Limb Praxis and

Apraxia proposed by Rothi, Ochipa and Heilman (1991, 1997). The Cognitive

Neuropsychological Model of Limb Praxis and Apraxia was derived from the study of

brain injured patients (Rothi et al., 1991). In order to systematically study errors

produced during pragmatic actions, a theoretical model is needed that explains the

relationship between frontal apraxia, conceptual/ ideational apraxia, and ideomotor

apraxia and incorporates action information from the Unified Hypothesis (Schwartz &

Buxbaum, 1997) and movement information from the Cognitive Neuropsychological

Model of Limb Praxis and Apraxia (Rothi et al., 1992, 1997). The resulting combined

model can then be used to make predictions about pragmatic errors which can be

subsequently studied. The purpose of the present study is to describe a theoretical

model of pragmatic action based on Schwartz and Buxbaum (1997) and Rothi et al.

(1997) which is then used as a framework to examine the nature of movement and

action errors produced by brain-injured patients. Pertinent literature is reviewed

relating to the genesis and growth of the study of movements and actions.

Review of the Literature

Historical Perspectives and Theoretical Models of Apraxia

Steinthal, in the late 1800s, was the first to use the term apraxia in describing a

disturbance in skilled limb movements as the result of brain damage (cited in Hecaen &

Rondot, 1985). Steinthal wrote that apraxia consisted of a disturbance in the

relationship between movements and the objects upon which the movements were

enacted (cited in Hecaen & Rondot, 1985). Subsequently, there was a lack of

consensus in discussions of the mechanism for and definition of apraxia. For example,








other researchers had observed the disturbances of object related movements in aphasic

patients but attributed both deficits to asymbolia, a generalized disturbance in the

comprehension or production of symbols in any modality, including language and

gesture (Finkelnstein, cited in Duffy & Liles, 1979; Critchley, 1939). Goldstein (1948)

also related disorders of action to the patient's aphasia and included skilled movement

problems within a definition of aphasia. Pick (1905), however, defined apraxia as an

asymbolia which was not included within a definition of aphasia. Another mechanism

proposed to explain apraxia was posited by Kussmaul (cited in Hecaen & Rondot,

1985) who defined apraxia as an agnosia, an impairment in the recognition of tools,

which then affects the movements produced with tools. In 1905, Liepmann described a

patient, who presented with a severe inability to produce volitional movements with the

left hand as well as profound aphasia. Liepmann's point in describing this case was that

a disorder of language or gnosis could not explain apraxia of only one hand.

Movement failures created by language or gnosis deficits would affect both hands.

Thus, Liepmann (1900/1977, 1905a/1980a, 1905b/1980b, 1907/1980c, 1920) was the

first to describe the mechanism of apraxia as a disorder of movement planning.

Liepmann (1905b/1980b) studied 89 brain-damaged patients, 42 with left

hemiplegia (thus suspected to have right hemisphere lesions), 41 with right hemiplegia

(thus suspected to have left hemisphere lesions), 5 non-hemiplegic with aphasia (left

hemisphere lesions), and one who was neither hemiparetic or aphasic but was apraxic.

The patients were asked to produce three types of movements 1) expressive

movements such as waving and saluting; 2) transitive and intransitive movements to

command (from memory) such as playing an organ grinder and snapping the fingers;








and 3) manipulations of actual tools such as combing hair with a comb and writing with

a pen. Liepmann found that the patients with right hemisphere damage rarely made

errors on these tasks whereas the patients with left hemisphere damage made frequent

errors. Within the groups of patients with left hemisphere damage, approximately half

showed evidence of apraxia and of these, twenty-five percent showed impairments

when manipulating the actual tools (Liepmann, 1905b/1980b). Based upon these

observations, Liepmann proposed that the left hemisphere, specifically the parietal

region, was responsible for the skilled production of both hands (Liepmann,

1905b/1980b). He argued that the right hemisphere is dependent upon the plans and

directives of the left hemisphere for learned movement and that the right hemisphere

receives movement planning information from the left hemisphere via the corpus

callosum. Liepmann proposed the existence of movement formulae which he defined

as knowledge of the course of action (time-space sequences) required to complete an

action goal as well as the semantic information about the tool and object used. The

movement formulae may be implemented by retrieval of innervatory patterns

(configurations of neural connections specialized for particular movement patterns)

which communicate directly with the motor system for movement production (Figure

1-1) (Liepmann, 1905b/1980b).

Additionally, further support for Liepmann's proposal that the left hemisphere

was responsible for the skilled movements of both hands was found in the case

described by Liepmann and Maas (1907) of a patient with a lesion of the corpus

callosum who was unable to produce skilled movements with his nonparalyzed left








Visual/Gestural Input


Visual Analysis


SMovement Formula I


I Innervatory Patterns


Left Right
Primary Primary
Motor Motor
System System
/ \


Right Hand
Gestural
Production


Left Hand
Gestural
Production


Figure 1-1. Liepmann's model of praxis (taken from Rothi, Ochipa, & Heilman, 1991).








hand. He was unable to write and could not even bring his hand into the writing

position. The patient also showed deficits in using actual tools/objects. Liepmann

hypothesized that the effect of corpus callosum lesion in this case was to disconnect the

movement formulae of the left hemisphere from the primary motor cortex of the right

hemisphere--lesion A in Figure 1-1 (Liepmann & Maas, 1907).

Subsequently, Liepmann (1905a/1980a, 1905b/1980b) described three subtypes

of apraxia--limb kinetic apraxia, ideo-kinetic or motor apraxia, and ideational apraxia.

Limb kinetic apraxia was described as a loss of the kinetic components of engrams

resulting in coarse or unrefined movements with movements that no longer have the

appearance of being practiced over time. Ideo-kinetic or ideomotor apraxia was

described as a loss of the voluntary ability to perform learned movements. Ideational

apraxia was described as an impairment of ideational (conceptual) knowledge resulting

in loss of the conceptual linkage between tools and their respective actions as well as

the ability to sequence correctly produced movements (Liepmann, 1905a/1980a,

1905b/1980b). Integral movements may be left out of a series, produced in the wrong

order, or correct movements may be produced with the wrong tools (cited in Brown,

1988). By describing these praxis subtypes, Liepmann (1905a/1980a, 1905b/1980b)

proposed that praxis is supported by a multicomponential system that can be

differentially impaired.

Liepmann's proposals were resurrected by Geschwind in 1965 who discussed

them in the light of new human and animal data. Geschwind supported the notion of

the dominance of the left hemisphere for learned movement skill and of the existence of

movement formulae or memories (stored representations). He also supported








Liepmann's statements that apraxia resulted from lesions in the dominant parietal

region but argued that the important area was the underlying white matter pathways,

specifically the arcuate fasciculus, and not the cortex.

Geschwind proposed that in the left hemisphere, the white matter pathway that

runs deep to the supramarginal gyrus connecting the visual association areas and speech

areas to the frontal lobes is the pathway by which motor responses are carried out in

response to verbal and visual stimuli. These pathways terminate in motor association

area, Brodman's area 6 (Geschwind, 1965). Movement information processed by area 6

is then relayed to the ipsilateral primary motor cortex, Brodman's area 4, as well as the

contralateral motor association area, area 6. Movement information processed by

contralateral area 6 is then relayed to it's affiliated primary motor cortex, area 4. (See

Figure 1-2.)

Based upon this neural mechanism, Geschwind proposed that the apraxia of the

left arm in a patient with right hemiplegia is not caused by the same lesion that caused

the hemiplegia. Rather, the apraxia results from coincidental damage to either area 6,

the callosal fibers connecting the left area 6 and the right area 6, or to the projections to

area 6 from the parietal lobe. This argument implied that a lesion anterior to the left

motor cortex which affects area 6 but spares area 4 should result in a bilateral apraxia.

However, Geschwind stated that as of his writing, no such cases have been described

presumably because lesions of left area 6 would most likely also affect area 4 due to

their anatomical contiguity.






































Figure 1-2. Geschwind's model of praxis. Lateral view of the left hemisphere.
AF = arcuate fasciculus; SMA = supplemental motor area or motor
association cortex; MC=primary motor cortex; VAC = visual
association cortex; VC = primary visual cortex. The arrows indicate
major connections of the areas shown (from Heilman & Rothi, 1985).








Assuming that visuokinesthetic engrams (Liepmann's movement formulae) of

the left parietal lobe are important for correct production of gesture as well as critical to

the decoding of seen, familiar gestures, Heilman, Rothi, & Valenstein (1982) postulated

that damage to the visuokinesthetic engrams would result in difficulty with both the

comprehension and production of gesture. The authors argued that measuring a

patient's ability to discriminate well-performed from poorly performed movements may

enable the clinician to determine whether or not a patient's apraxia resulted from

destruction of these engrams through lesions involving posterior cortical regions:

specifically left parietal lobe. An apraxic patient with preserved visuokinesthetic

engrams would be able to discriminate between good and poor performance while an

apraxic patient with damage to the visuokinesthetic engrams would not be able to make

that discrimination. The authors tested this theory by giving 20 apraxic subjects a

gesture discrimination test. The examiner named a target action and the subjects were

asked to watch a videotape of gesture performances and choose the performance out of

three possible for each command that best represented the named target. Patients with

more difficulty discriminating the gesture productions had posterior lesions (and

presumably damaged visuokinesthetic engrams.) Based on these results, Heilman et al.

(1982) discussed the possibility of two types of ideomotor apraxia: one type which

presents with poor production and comprehension of gestures and results from damage

to the left supramarginal or angular gyrus in the parietal lobe and a second type which

presents with praxis production problems alone and results from damage that does not

involve this parietal region. (See Figure 1-3.)

















































Figure 1-3. Heilman and Rothi's Model of Praxis. View from top of brain.
W = Wernicke's area; VA = primary visual area; VAA = visual association
area; AG = angular gyrus; SMG = supramarginal gyrus; PM = premotor
area (motor association cortex); M = motor cortex; CC = corpus callosum;
LH = left hemisphere; RH = right hemisphere. The arrows indicate major
connections of the areas shown (from Heilman & Rothi, 1985).








Cognitive Neuropsvchological Model of Limb Praxis and Apraxia

Recognizing Liepmann's assertion that skilled praxis was the product of a

complex multi-component system and also recognizing that numerous cases were

described in the literature that reflected behavioral fractionation of praxis related

behavior, Rothi, Ochipa, and Heilman (1991, 1997) proposed a cognitive neuro-

psychological model of apraxia in an attempt to capture these dissociations. A

schematic representation of this model can be found in Figure 1-4. In this model, the

term "action-lexicon" is used as the gestural equivalent to the term "lexicon" used in

language which distinguishes that part of the language system which gives a processing

advantage for words that the person has previously experienced (Rothi & Heilman,

1985). Therefore, the action-lexicon was defined as that part of the praxis system

which gives a processing advantage for movements that the person has previously

produced (Rothi et al., 1991, 1997). That is, it is an action memory. The action lexicon

was divided into input and output components to account for patients who demonstrate

spared gesture comprehension with impaired imitation and gesture to command. The

authors stated that for these patients, spoken language gains access to the output action

lexicon via semantics without being processed by the input action lexicon (Figure 1-5).

In contrast, deficits in gesturing to command with spared repetition and spared

comprehension were explained by dysfunction at or after the output action lexicon

while sparing the innervatory patterns (Rothi et al. 1991).

DeRenzi, Faglioni, and Sorgato (1982) and Rothi, Mack, and Heilman (1986)

discussed cases that provide evidence for modality-specific apraxic deficits. That is,

the praxis performance of the reported patients differed as a result of input modality











Auditory/Verbal Visual/Object Visual/Gestural
Input Input Input


Figure 1-4. Cognitive Neuropsychological Model of Apraxia (from Rothi, Ochipa, & Heilman, 1991).














AuditoryNerbal Visual/Object Visual/Gestural
Innt Input Input


Figure 1-5. Delineation of Gesture to Command Pathway (from Rothi, Ochipa, & Heilman, 1991).








(e.g. auditory, tactile, or visual). Rothi et al. (1991, 1997) accounted for the differences

in performance as a result of input modality by incorporating separate input systems for

visually presented gestural information, tactily presented (no visual input), and

auditorily presented verbal information. For visual/gestural information, the

product of visual analyses accesses either the input action lexicon or the innervatory

patterns directly (Figure 1-4). The presence of this direct route from visual analysis to

innervatory patterns may account for those patients who can imitate gestures but can

not discriminate or comprehend gestures (Rothi et al., 1986). That is, in cases where

gesture imitation is not possible through input to output lexicons or input lexicon to

semantics to output lexicon, gesture imitation may occur through a "non-lexical" route

similar to the route in which nonwords are imitated in language. For visual/ object

information, visual analyses activate the object recognition system which produces a

structural description of the viewed object. It is hypothesized that the structural

descriptions then access either the semantic system or action output lexicon. The

connection between the visual analyses and the action output lexicon may allow for the

production of gesture from the visual modality without accessing semantics. Visual

object information being processed accesses the action output lexicon without prior

processing by the action input lexicon because the action output lexicon pertains to

codes of the "to-be-performed" actions while the action input lexicon pertains to codes

of "perceived" actions (Rothi et al., 1997). This route (visual analysis to action output

lexicon) may account for the ability of certain patients who can not comprehend the

meaning of the gestures because of a failure of the action semantic system, but can

accurately produce gestures to shown tools. These patients may have intact








comprehension of objects but impaired comprehension of actions thereby suggesting a

dissociation within the semantic system for information about objects and information

about actions. Auditory/verbal input is transmitted to auditory association areas that

perform auditory analyses and activate the phonological input lexicon. Information

processed by the action input lexicon can be analyzed semantically by the semantic

system or phonologically for output by the verbal output lexicon. In figure 1-4, the line

connecting the region where auditory analysis is performed to the phonological buffer

may be the route that accounts for those patients who can repeat but not comprehend

verbal information (Rothi et al., 1991).

Roy and Square (1985) propose that praxis processing involves not only

information about gestural production but conceptual information as well. Within the

Rothi et al., (1991, 1997) model, conceptual analysis of praxis is accomplished by the

action semantic component. Roy and Square (1985) and Raymer and Ochipa (1997)

describe praxis conceptual knowledge as knowledge of the functions of tools and

objects, knowledge of actions not related to tools and objects, and knowledge of

combining actions into sequences. Rothi et al., (1991, 1997) state that actions depend

upon the interaction between the conceptual knowledge described above and the

sensorimotor information contained in the motor programs. The action semantic

system contains the conceptual knowledge relating to tools, objects, and actions. In

figure 1-4, the line in the semantics area accommodates recent information that action

semantics can be compartmentalized from nonaction semantics. Support for this

separation within semantics is found in reports of patients with optic apraxia who were

able to accurately gesture to command but were unable to accurately imitate the same








gestures or gesture to visually presented tools (Raymer, Greenwald, Richardson, Rothi,

& Heilman, 1992) as well as studies of Alzheimer's patients with impaired action

semantics with spared nonaction semantics and vice versa (Ochipa et al., 1992,

Raymer, 1992). This separation also accommodates those researchers who claim that

there are multiple semantic systems which reflect the modality and nature of input

material (Beauvois & Saillant, 1985; Paivio, 1986; Shallice, 1987, 1988).

By comparing patients' performance across tasks, researchers and clinicians

may use this neuropsychological model to ascertain the nature of the processing

damage and make predictions regarding performance. For example, when gesturing to

command, patients would receive auditory/verbal information (e.g. "show me how to

use a hammer") which would be processed phonologically by the phonological input

lexicon followed by the conceptual analysis of the action semantic system. After

semantic analysis, the information would be processed by the action output lexicon and

the relevant innervatory patterns would be accessed for production of the gesture by the

motor system (Figure 1-5). When gesturing to a visually presented tool/object, patients

would visually analyze the tool/object and corresponding information would be utilized

by the object recognition system with subsequent access to semantics, action output

lexicon, and the innervatory patterns for production by the motor system (Figure 1-6).

According to the model in Figure 1-6, patients must access semantics to gesture to

command correctly but do not have to access semantics to gesture to a visually

presented tool/object. Therefore, by comparing the patient's performance in these two

tasks, researchers and clinicians may be able to evaluate the integrity of the semantic

system.














AuditoryNerbal


Visual/Object
Input


Visual/Gestural


Figure 1-6. Delineation of Gesture to Visually Presented Object Pathway (from Rothi, Ochipa, & Heilman, 1991).








Right Hemisphere Contributions to Praxis

In addition to his proposals that the left hemisphere was dominant for learned,

skilled movement, Liepmann also acknowledged that the right hemisphere may also

have the ability to plan and generate skilled movement (Liepmann, 1905a/1980a,

1907/1980c, 1920).

However, Liepmann considered the right hemisphere to be subordinate to the

dominate left hemisphere and would require the activation of the movement formulae

(visuokinesthetic engrams) of the left hemisphere for correct production of skilled

movement in situations where the movements were done without manipulation of the

actual tool and object. That is, the right hemisphere depends on the integrity of the left

praxis system for production of learned, skilled movement in the absence of visual or

tactile information. In other words, the right hemisphere relies on the intact left praxis

system to produce correct actions when tools and objects are not used (Liepmann,

1905a/1980a, 1920; Maher, 1995).

This ability of the right hemisphere to produce correct movements and actions

in context with actual tools and objects was confirmed by Rapcsak, Ochipa, Beeson,

and Rubens (1993) who reported a patient, G.K., who suffered destruction of the entire

left hemisphere following a massive stroke. G.K. was strongly right-handed and it

could be hypothesized that the left hemisphere was dominate for language and praxis.

However, results of complete and intensive praxis testing which evaluated each input

and output modality for production and comprehension of learned, skilled movement

indicated that G.K. was severely impaired in all praxis tasks when actual tools and/or

objects were not used. In fact, G.K. reported no difficulty performing common








everyday tasks such as preparing meals and fixing household items in which tools and

objects are used together and in the relevant context.

Maher (1995) studied the role of the right hemisphere in production of learned,

skilled movements by evaluating the temporal and spatial aspects of gestures produced

by patients with left or right hemisphere brain lesions and matched normal control

subjects using a scoring system described by Rothi, Mack, Verfaellie, Brown, and

Heilman (1988). The subjects produced gestures to auditory command using the arm

and hand ipsilateral to the side of lesion to control for possible effects of hemiparesis.

The gestures were scored by two independent raters who were blind to the nature of the

study. The raters received intensive training before each scoring session which resulted

in high inter- and intra-rater reliability. The raters first scored the gestures pass/fail

according to correctness. Then, if a gesture was scored as incorrect, error type(s) were

awarded based on the above error classification system which included the following

error types: internal configuration, external configuration, occurrence, amplitude,

movement, and sequencing of movements (e.g., to use a key one must insert, turn, and

remove the key in that order). Results of the study showed that the left brain lesioned

subjects performed significantly worse than both the right brain lesioned subjects and

normal control subjects overall. When frequency of each error type was compared

across groups, there were significant differences for all error types except movement

and amplitude. Pairwise comparisons were then performed to distinguish etiology of

the differences. These comparisons suggested that the significant difference between

the left and right brain lesioned subjects for movement errors can be accounted for by

extremely poor performance by the left brain lesioned subjects when compared to the








other subject groups. This was a significant result because it supports the notion that

the left hemisphere is dominate for praxis. Apraxia is inherently a disorder of learned,

skilled movement and significant difficulty in the movement aspects of skilled

movements should be observed following damage to this system. The pairwise

comparisons also suggested that the significant difference for amplitude apects of

movement occurred between the right brain lesioned subjects and the normal controls

with no significant difference between the right and left brain lesioned groups. Maher

(1995) proposed that this result suggests that both the right and left hemispheres could

code spatial aspects of movement.

The Ecology of Apraxia

Significance to clinicians and patients

Other than its lateralizing value, apraxia has not been considered of significant

practical importance. DeRenzi (1985) said that apraxia rarely appears in everyday

situations and only emerges "out of context as a purposeful response to an artificial

request" (page 134.) Because ideomotor apraxia was considered to be an examination-

bound symptom that would not attract a clinician's attention unless the clinician

specifically looked for it, apraxia may have been overlooked and underestimated in

brain damaged patients (DeRenzi, 1985).

Many clinicians might agree that patients do not spontaneously complain of

apraxic disturbances (Heilman & Rothi, 1985). One possible explanation for this is that

apraxic patients may have anosognosia (unawareness of the disorder) for their praxis

difficulties (Rothi et aL., 1986). Another possible explanation is that patients with

apraxia frequently have a co-occurring right hemiparesis. Therefore, patients may








associate their clumsiness to use of the non-dominant left hand for tasks in which they

previously used their dominant right hand. While apraxia may be present, praxis

performance may improve when manipulating actual objects or tools (Geschwind,

1965) and caretakers or family members, who rarely have the opportunity to observe

tool/object use pantomime, may not be aware of the disorder. In fact, many family

members may provide assistance in self-care skills such as feeding and dressing for

convenience and efficiency purposes (Heinemann, Roth, Ciehowski, & Betts, 1987).

In a hospital or rehabilitation center, it is possible that the environment of the

patients may be controlled and geared toward partial functional independence where a

patient can complete at least part of a task independently. For example, in a nursing

home environment it has been observed that when it is time for a patient to brush

his/her teeth, the nurse may put the toothpaste on the toothbrush, hand the toothbrush to

the patient, and tell the patient to "brush your teeth." It has also been observed that

during mealtime, the nurse may prepare the food for the patient by cutting the meat,

putting the straw in the drink and/or handing the correct utensil to the patient. A study

by Heinemann and colleagues (1987), which examined functional outcome of stroke

rehabilitation programs, found that all of the independence gained during rehabilitation

in the area of feeding was lost by three months after discharge. They proposed that loss

of a controlled environment that promoted functional independence (e.g., the caregiver

no longer prepared the environment to maximize the patient's skills, but began to assist

the patient directly) may explain this decline in function.

Another study examined factors that influenced a person's ability to return to

work after a stroke (Saeki, Ogata, Okubo, Takahashi, & Hoshuyama, 1993). The








authors followed 244 patients, ages 24 to 65, who were actively employed at the time of

their strokes and extracted information from the admission medical records of each

patient. Results indicated that 58% of the patients returned to work by the time of the

follow-up which varied from 8 to 77 months post-stroke. The most important factor

that determined a patient's ability to return to work was severity of muscle weakness on

admission. The second most important factor was presence of apraxia. The odds of a

patient without apraxia returning to work were determined to be four to five times

greater than for patients with apraxia (Saeki et al., 1993). However, possible

limitations of this study are that the authors combined ideomotor apraxia with

ideational apraxia, dressing apraxia, and constructional apraxia in the analyses; none of

which were operationally defined. As a result, it can not be determined which one or

more types of apraxia affected the patients' ability to return to work or whether the

authors' definition of apraxia was similar to or different from previously published

definitions.

Apraxia in natural contexts

The first study to document the effects of apraxia in the non-clinical

environment was conducted by Sundet, Finset, and Reinvang (1988). These

researchers investigated variables that affected a patient's ability to function in the

home environment following discharge from a rehabilitation hospital. They sent

questionnaires to 68 left hemisphere stroke patients and 77 right hemisphere stroke

patients six months after discharge from rehabilitation. The questionnaire focused upon

activities of daily living and the amount of "dependency" the patients displayed in the

home environment. Dependency was defined as the increased need for caregiver








assistance in performing tasks of daily living. The patients were asked 13 simply

written yes/no questions such as "requires use of kitchen aids" and "requires help to

dress" (Sundet et al., 1988, p.369). Caretakers were asked to answer the questions in

the event that a patient could not. Sundet and colleagues compared the relationship

between results of the questionnaire (measure of dependency) and neuropsychological

deficits such as hemiplegia, aphasia, nonverbal memory deficits, neglect, and apraxia;

the presence of which were determined by review of each patients' medical records.

Results of the study indicated that the highest predictor of dependency for the left

hemisphere damaged subjects was the presence of apraxia.

Limitations of the Sundet et al. (1988) study include the authors' reliance upon a

questionnaire filled out by caregivers rather than direct patient observation to determine

the patients' ability to function independently in a natural environment and context.

There was also a six month hiatus between the neuropsychological testing and

determination of the patients' "dependency" during which time the patients'

performance on the neuropsychological tests (and, in turn, the presence of

neuropsychological deficits including apraxia they were reported to document) may

have changed. Although a study by Rothi and Heilman (1985) documented that over

80% of patients who are acutely apraxic remain apraxic six months later, it is unknown

how many of the 68 left hemisphere damaged subjects in the Sundet et al. (1988) study

would have been considered apraxic at the time of dependency determination.

One study in which patients were directly observed and which did not have a

hiatus between clinical examination and experimental procedures was conducted by

Foundas, Macauley, Raymer, Maher, Rothi, and Heilman (1995). Foundas and








coworkers (1995) examined actual tool use during mealtime in 10 left hemisphere

stroke patients and 10 neurologically normal controls matched for handedness, age,

gender, and education. Subjects were tested for presence of aphasia and apraxia and

then, on the same day as testing or no more than three days later, were videotaped while

eating a meal. For each subject the meal tray was placed on a table in front of the

subject and foil items such as a toothbrush, comb, and pen were included on the tray in

addition to the standard tools and utensils for eating (spoon, fork, knife, napkin, and

condiments). The location of the standard tools and the foils were counterbalanced to

control for their left to right placement on the tray. The examiner left the room and no

assistance was given to the subject in preparing or eating the food.

Results of the Foundas et al. (1995) study documented two main differences

between the control and experimental subjects. The first difference became apparent

when the stages of eating (preparatory, eating, and clean-up) were compared across

groups. During the preparatory phase, behaviors included such things as opening

condiment packages, placing the napkin on the lap, cutting the meat, and putting sugar

in the tea. During the eating phase, the food was eaten. During the clean-up phase,

behaviors such as putting the napkin back on the tray, putting the utensils on the plate,

and pushing the tray away were accomplished. The first difference between the groups

was that 80% of the control subjects proceeded through all three phases of the meal

compared to only 20% of the experimental subjects. The authors reported that the

control subjects respected the boundaries of each phase whereas the experimental

subjects did not. That is, the control subjects showed clear, distinct beginning and end

points of each phase while the experimental subjects tended to have preparatory, eating,








and clean-up actions interspersed randomly throughout the meal as if there was a lack

of anticipation for pending task demands. The authors also reported that within the

eating phase, the control subjects ate using eating patterns that were idiosyncratic

within subjects but consistently developed across subjects. For example, a control

subject might eat portions of the main course (e.g., meat) followed by the bread, the

vegetables, and drink of tea and repeat this pattern over again until all food had been

eaten. In contrast, the experimental subjects tended to eat in a more random fashion

with no definable eating pattern.

In the Foundas et al. (1995) study, the experimental subjects produced more

"tool errors" (errors of action in using tools such as spoon, and fork) than "nontool

errors" (errors of actions that do not involve tools such as wiping the face or moving

the glass) while the control subjects did not produce any errors. One might suggest,

however, that because the experimental subjects had a hemiparesis and the comparison

group did not, the differences between the groups may be related to the presence of a

primary motor defect in the apraxic group. However, an underlying motor system

deficit could not explain the difference in eating praxis performance in these two

groups because a motor deficit should have affected both tool and nontool actions

equally, and it did not. Therefore, even though the experimental subjects had a right

hemiparesis, the action errors they produced, selective only for praxis related to tool

use, were not the result of an underlying motor problem.

Results of the Foundas et al. (1995) study indicated that while the overall eating

time of the experimental subjects did not differ from the control subjects, the

experimental subjects made fewer actions in general, used fewer utensils, and often








misjudged the advantage of using a tool. The experimental subjects produced

significantly more incorrect actions when compared to the controls and as a result were

less efficient in executing individual tool actions. Tool misuse and mis-selection errors

were produced by all but one experimental subject who happened to score within the

normal range on the test for ideomotor apraxia. Errors such as eating ice cream with a

fork (mis-selection), cutting with a spoon (misuse), stirring with a knife (misuse), and

wiping one's face with a slice of bread (mis-selection) were observed. A schematic

representation of misuse and mis-selection errors can be found in Figure 1-7. In

contrast, the normal control subjects made more actions in general, used more utensils,

and used the utensils correctly throughout the meal. The tool actions of the

experimental subjects were often incomplete or imprecise so that a desired goal was not

achieved. Foundaset al. (1995) argued that the high correlation between number of

action errors produced and presence of ideomotor apraxia suggested that ideomotor

apraxia caused the production of action errors in this natural context.

A series of studies by Schwartz and colleagues also examined action errors in

activities of daily living by brain damaged subjects. The first study by Mayer, Reed,

Schwartz, Montgomery, and Palmer (1990) examined a group of brain-injured patients

in a rehabilitation hospital. Forty-five patients with traumatic brain injury and six

patients with stroke were observed during activities of daily living such as brushing

teeth and making coffee. The authors found that 49% of the patients produced errors of

action during activities of daily living. Unfortunately, due to the nature of the brain

damage of the patients they studied, it is not known whether the action errors were the

result of the















Misuse


object


Mis-selection


object


action


object


Misuse and Mis-selection


object


action


object


Figure 1-7. Schematic Representation of Misuse and Mis-selection Action Errors.
// indicate location of errorss.


action


action


object


action


action


object


action








general intellectual impairment or executive function defects commonly found in

traumatic brain injured patients or of apraxia.

Mayer et al. (1990) developed a theory of action based upon Norman and

Shallice's (1986) model of attentional control of action which proposes two modes of

control of action, one which is automatic and one which requires deliberate attentional

control. Mayer et al. (1990) studied the action productions within a natural context of

45 patients with closed head injury and proposed that impairment of the deliberate

attentional control route was the genesis for production of action errors in their patients

with brain damage. For this study, an action coding system was developed that

describes the sequence of actions produced by brain damaged patients as either A-I

units (general actions, not task-specific such as rinsing a glass), A-2 units (specific,

task-related actions such as unscrewing the cap from a toothpaste container), or crux A-

1 units (the action that accomplishes the next goal such as squeezing the toothpaste

onto the toothbrush when preparing to brush teeth). The action coding system enabled

the researchers to examine which actions were produced out of sequence, were omitted,

or were perseverated by applying the action coding system to a script of the patient's

actions and examining the relationship and sequence of the A-I, A-2, and crux A-I

units in the patient's production to that of a prototype script.

Mayer et al. (1990) then applied the action coding system to a script of actions

produced by a brain injured patient during a task of daily living. One patient, H.H., had

suffered a bifrontal injury with resulting aphasia and callosal apraxia in which he was

apraxic with his non-dominant left hand and not apraxic with his preferred right hand.

H.H. was videotaped during breakfast and a script of his actions was written. The








action coding system was applied to the script to analyze action sequences. The authors

reported that even though H.H. produced frequent errors such as pouring tomato juice

on waffles and attempting to spoon oatmeal into tea, his main deficit was in the

efficiency of his action plans. That is, A-1, A-2, and crux A-I units were produced

randomly and not in a logical order or sequence to accomplish the goal of the task at

hand. In addition, there was the appearance of action errors that included misuse and

mis-selection of tools as well as movement errors that included sequencing and timing

of movements. The authors argue that these errors, misuse and mis-selection of tools as

well as sequence, were not a result of the patient's apraxia because "in the majority of

patients with documented apraxia, there (was) no functional consequence of the apraxic

impairment" (Mayer et al., 1990, p.280). Foundas and colleagues, however, would

argue that the above statement is not an accurate assumption in that the authors did not

test for apraxia and as a result could not speak towards the significance of apraxia to the

production of action errors in their patients.

Schwartz, Reed, Montgomery, Palmer, and Mayer (1991) continued their work

by applying the action coding system to two specific tasks--making coffee and brushing

teeth--as produced by H.H., the brain damaged patient described by Mayer et al.

(1990). H.H. was videotaped performing these tasks in the natural environment and

scripts of his actions were analyzed according to the action coding system described by

Mayer, et al. (1990). The authors noted that H.H. had a higher susceptibility to object

substitutions and object misuse as well as a variability of action errors across time.

Regarding normal action processing, the authors argue that learned actions are not








executed as whole programs but are organized into hierarchies of temporally structured

units that depend on the prefrontal area of the brain for integration and production.

Based on the results of the study described above, Schwartz et al. (1991)

proposed four theoretical foundations of intentional actions and action errors: the first

tenet claims that intentional action involves activation of an action plan; the second

claims that the planning and execution of intentional actions are closely coupled in real

time; the third claims that intentional actions are integral to all purposeful behavior; and

the fourth claims that errors of action occur due to weakening of the top-down

formulation of action plans. This weakening of the action plans allows the patient to be

influenced by irrelevant objects and actions leading to a susceptibility of object related

errors of action. Schwartz and colleagues (1991) also proposed that the condition of

action disorganization observed in brain injured patients while performing activities of

daily living be termed frontal apraxia.

The explanation of the mechanism of frontal apraxia offered by Schwartz et al.

(1991) is not contradictory to Liepmann (1900/1977,1905a/1980a, 1905b/1980b, 1907/

1980c, 1920) and Heilman and Rothi's (1985, 1993) description of movement

memories. Rather, Liepmann, Heilman and Rothi's proposals focus upon a retrieval

system for the memories of unique learned skilled movements that span from single

discreet actions (the action lexicons) to subcomponents (innervatory patterns) of these

actions. In contrast, the proposal of Schwartz et al. (1991) focuses upon the internal

relationship between discreet component actions of larger action goals. Both Liepmann

(1920) and Schwartz et al. (1991) use the term "frontal apraxia" to describe the

movement/action errors that occur when the internal relationship between discreet








action representations in the larger action goal context have been loosened or

disorganized by brain damage.

Schwartz, Montgomery, DeSalme, Ochipa, Coslett, and Mayer (1995) expanded

the study of action errors by examining J.K., a patient with traumatic brain injury,

whose behavior was similar to the patient (H.H.) studied by Mayer et al. (1990) and

Schwartz et al. (1991). J.K. was videotaped while eating breakfast and brushing his

teeth and the tapes were scored according to the action coding system described by

Mayer et al. (1990). J.K. also underwent a battery ofneuropsychological tests that

included object identification and recognition, conceptual/semantic knowledge,

functional/use knowledge, gesture knowledge, gesture production, other language

testing, and memory testing. Results indicated that J.K. evinced numerous action errors

during the activities of daily living, especially tool misuse errors. The authors

postulated that the tool errors were not due to a conceptual/ideational apraxia because

J.K. was able to correctly identify tools by name and by function as well as match tools

with their corresponding object. However, the authors argued that J.K.'s tool misuse

errors were not the result of a weakened top-down processing of action plans found

with frontal apraxia because J.K. was able to demonstrate or gesture the appropriate use

for a tool from visual and tactile cues, a high level task (Schwartz et al., 1995). The

authors concluded that in order to accommodate J.K.'s full range of performance, both a

top-down and a bottom-up impairment of intentional action planning should be adopted

(Schwartz et al., 1992). That is, the loosening or disorganization of relationships

between discreet action representations could be the result of damage to either higher or

lower order mechanisms within the action planning and movement production system.








Schwartz et al. (1993) adapted the action-trigger-schema (ATS) framework

developed by Norman and Shallice (1986) as a means of organizing actions and

classifying action errors. The ATS framework proposes that action plans are composed

of schemas which are organized memory structures that integrate different types of

information for use with actions. These schemas and their interrelations are developed

over time through individual experiences and can generate positive or negative

activation to specialized systems such as the motor system from the action plans.

Familiar action sequences such as eating are represented by groups of schemas

organized hierarchically. High level schemas occur in a top-down fashion while the

low level schemas occur in an orderly fashion according to the logical progression of

actions within the schema. For example, the higher level schema for eating breakfast

would be the intent to eat and complete the meal while the lower level schemas would

be those actions required to eat the meal such as opening a juice carton, picking up a

fork, and taking a bite of food. Using the ATS framework, the authors proposed that

action errors are due to the loss or instability of activation within the action schema

network possibly due to a weakening of the connections among schemas (Schwartz et

al., 1993).

Schwartz et al. (1993) argue that a patient with frontal apraxia will produce

errors differently during an activity of daily living than a patient with a disorder of

attention. Specifically, a frontal apraxic patient with dysfunction in the ability to plan

and coordinate action sequences may exhibit incoherence and intrusions of actions

which result in a fragmentation of behavior. Patients with frontal apraxia may also use

tools and objects in novel or bizarre ways because the action plan for using the tools








and objects is inaccurate, nonspecific, and faulty. In contrast, an attention disordered

patient may exhibit intrusions of irrelevant actions and temporary derailments to other

tasks resulting in inefficient but coherent action sequences.

Schwartz et al. (1993) addressed the issue of whether the observed action errors

were related to an ideational apraxia by comparing the action performance of H.H., a

traumatic brain injured patient described in Schwartz et al. (1991), to the action

performance of a left-handed, right hemisphere damaged patient described by Ochipa

et al. (1989) who displayed ideational (conceptual) apraxia. Although both patients

produced numerous errors of action during activities of daily living, H.H. evinced

misuse of tools with spared tool function knowledge while the Ochipa et al. (1989)

patient evinced spared production with impaired tool function knowledge. Schwartz et

al. (1993) posited that frontal apraxia and ideational apraxia have different underlying

mechanisms in that patients with ideational apraxia have impaired conceptual-semantic

knowledge while patients with frontal apraxia have spared conceptual-semantic

knowledge.

Mayer et al. (1990) and Schwartz et al. (1991, 1993, 1995) also applied the

action coding system to the performances of a subject who suffered a hemorrhagic

stroke (J.H., discussed in Schwartz et al., 1991) and a subject who suffered diffuse

damage following traumatic brain injury (J.K., discussed in Schwartz et al., 1993,

1995) during two activities of daily living. Although the errors produced by the

patients were described, the nature of the patients' brain damage did not allow the

researchers to make conclusions about brain-behavior relationships. Both hemorrhagic

strokes and traumatic brain injury may be associated with more damage to the brain








than appears on CT/MRI scans. One other limitation of this series of studies is that

although it was not their intention to include ideomotor apraxia as a possible factor,

ideomotor apraxia is likely to have been present in their cases and may have been an

important factor.

Schwartz and Buxbaum (1997) proposed a "Unified Hypothesis" to explain the

errors of action observed in patients with frontal and/or ideational apraxia. The

author's schematic representation of this hypothesis can be found in Figure 1- 8. The

Unified Hypothesis is based upon the ATS framework which states that through

experience and learning, higher level schemas organize lower level schemas into

temporally ordered sequences. Experience and learning solidify connections between

higher and lower level schemas for specific tasks making the retrieval of the lower level

schemas automatic. For example, one higher level schema would contain information

on the sequence of actions required to eat a meal and would automatically activate

lower level schemas for cutting, drinking, wiping, and stirring, etc., as needed

throughout the meal process. Another high level schema would contain information on

the sequence of actions required to build and would activate lower level schemas for

getting the permit, hiring architects, buying materials, putting in the foundation, etc. In

turn, the schema for putting in the foundation would activate lower level schema for

marking the ground, digging the hole, pouring the concrete, and etc. Whether a schema

is considered "low" or "high" is determined by the relationship between schema within

the goal of the task. Schwartz and Buxbaum (1997) differentiated between learned,

routine actions and novel or nonroutine actions in that pre-existing connections existed

between high level and low level schemas for learned, routine actions but not for novel











Perpetual/Sensory Input


etion
ItenttEon::]


Figure 1-8. Schematic Interpretation of the Unified Hypothesis (from Schwartz & Buxbaum, 1997).


-"- Motor System








or nonroutine actions. The pre-existing connections between higher and low level

schemas allow learned, routine actions to become more automatic, but also more prone

to error (Schwartz & Buxbaum, 1997; Reason, 1990). According to Schwartz and

Buxbaum, (1997), it is the evocation of attentional factors at certain critical points

during activation of higher level schemas which opposes error tendencies and keeps the

error rate low.

Schwartz and Buxbaum (1997) discussed two key assumptions of the Unified

Hypothesis. The first key assumption is that damage to the movement engrams and/or

the perceptual systems, or the semantic systems that input to the movement engrams

would affect the ATS system by compromising its automatically. This lack of

automaticity leads to greater dependence on attention and specifically on supervisory

attention. Supervisory attention is evoked when automatic action routines are

insufficient or inappropriate to complete the task at hand (Norman & Shallice, 1980,

1986). Supervisory attention has been equated with working memory by Baddeley

(1986). Schwartz and Buxbaum (1997) proposed that "for the patient with faulty

access to gesture engrams or perceptual-semantic processing, successful performance

of familiar, pragmatic action, depends more heavily on [attentional] control processes"

(p.19).

The second key assumption of the Unified Hypothesis is that behavioral

consequences of this increased dependency on supervisory attention/working memory

depend upon a combination of the severity of damage to the ATS system and the status

of the supervisory attention/working memory, itself. That is, a deficit in one

component (activation of action schemas or attentional control systems) of the Unified








Hypothesis may not impact actions in natural contexts without some degree of damage

to the other component. The authors argue that it is this combination of deficits that

leads to action errors in natural contexts. That is, patients who have deficits in the

automatic activation of action schemas will not evince action errors unless there is also

a deficit in supervisory attention/ working memory and/or executive control processes.

Conversely, patients with deficits in executive control processes will not evince action

errors unless there are co-existing deficits in posterior processing systems which impact

the automaticity of the action schemas. In addition, patients with a cortical dementia or

diffuse head injury, may demonstrate the most serious errors of action during activities

of daily living because of the severity of deficits in both the automatic activation of

action schemas as well as executive control systems (Schwartz & Buxbaum, 1997).

The Relationship Between Apraxia and Pragmatic Action

It is possible that neither the model proposed by Rothi et al. (1991, 1997) or the

Unified Hypothesis proposed by Schwartz and Buxbaum (1997) can independently

account for all of the errors produced by brain-damaged patients during goal-directed

activities of daily living (pragmatic action). In the following discussion, pragmatic

action will be defined as the sequences of movements required to progress toward and

obtain a definable goal. The goal may be directed toward common, everyday actions

such as eating, or uncommon actions such as building a house (uncommon actions may

vary from person to person depending upon life experiences). In contrast, movement is

defined as changes in posture, place, or position which can occur within an action as

well as independently. Actions and movements have been disambiguated to represent

the notion that movements are the inherent motor aspect of performing actions (i.e.,








movements are the final common denominator for all actions) and that actions are a

series of movements arranged in hierarchically organized schemas according to the end

goal. Using these definitions of pragmatic action and movements, the Rothi et al.

(1991, 1997) Cognitive Neuropsychological Model of Limb Praxis and Apraxia

encompasses aspects related to the planning, organization, production and

comprehension of discreet movements (hereafter referred to simply as "movements")

while the Schwartz and Buxbaum (1997) Unified Hypothesis covers aspects related to

the planning, organization, production, and comprehension of actions which

incorporate discreet movements (hereafter referred to simply as "actions") and it is the

relationship or interaction between these two theories that is crucial for the

understanding of pragmatic action. In the following discussion, a relationship between

movements and actions will be proposed which will then be used as a foundation for

examining the production of errors during goal-directed activities of daily living.

It is proposed that higher level action schemas discussed by Schwartz and

Buxbaum (1997) access lower level schemas of movement organization found in the

action output lexicon as discussed by Rothi et al. (1991, 1997) (Figure 1-9). The action

output lexicon, hereafter referred to as praxicon as suggested by Heilman and Rothi

(1997), in turn, accesses the innervatory patterns and the motor system in that order, for

production of the movement (Rothi et al., 1991, 1997) (Figure 1-10). Within the high

level schemas, information about object, tool, sequence, and other aspects of action are

organized hierarchically into plans or scripts according to specific goals. It is also

proposed that the high level action schemas also contain attributes of movement

planning but under normal circumstances, the praxicon takes precedence.















High Level Schemas


Figure 1-9. Relationship between Actions and Movements.


Movements

(Rothietal., 1991,1997)


Actions

(Schwartz & Buxbaum, 1997)


Low Level Schemas











LOW LEVEL SCHEMAS


HIGH LEVEL SCHEMAS







Action

Schemas


Figure 1-10. Breakdown of Low Level Schemas.









According to Schwartz et al. (1991, 1993, 1995, 1997) there may be many

different levels of schema. That is, for one set of goal-directed action sequences, a

particular schema may be organized as a "lower level schema" whereas in a different

set, the same schema may be organized as a "higher level schema." In fact, it is

possible that schemas may be organized in such as way that a lower level schema may

be included in a higher level schema which is included in an even higher level schema

and etc. In order to conceptualize this framework it is proposed that within the high

level schematic system, information is stored in a distributed network framework and is

therefore subject to assumptions associated with parallel-distributed-processing (PDP)

models as described by Rumelhart and McClellan (1986). For example, three

assumptions about connections between nodes that are frequently activated together are

1) as the number of times of activation increases, the strength of the connection will

also increase, 2) as strength increases the nodes are activated faster, and 3) the stronger

the connections, the less likely they are to be degraded (Rumelhart & McClellan, 1986;

Nadeau. 1994). The corresponding assumptions for the high level schematic system are

1) stronger associations occur between informnation/schemas that are frequently

activated together, 2) the schemas with the most activation become learned, routine

actions and eventually become automatic, and 3) the schemas for common, learned

actions are less likely to show deficits following brain damage due to "a great deal of

redundancy [interconnectiveness] in the neural systems responsible for pragmatic

action" (Schwartz & Buxbaum, 1997, p2).

Norman and Shallice (1980, 1986) and Schwartz and Buxbaum (1997)

discussed the role of supervisory attention/working memory (SA/WM) in the








production of actions. The role of SA/WM was to monitor actions and step in as

needed during their production so errors would be circumvented. SA/WM also

contributed the most resources to monitor novel actions, a lesser amount of resources to

monitor non-routine actions, and the least amount of resources to monitor routine or

learned actions which could eventually become automatic and be produced correctly

without attentional resources (Schwartz & Buxbaum (1997). To account for these

different aspects of actions, the connection between executive functions (as defined by

Schwartz et al., 1991, 1993, 1995) and the output praxicon has been divided into four

routes, each requiring different amounts of attentional resources (Figure 1-11). It is

proposed that the route for production of automatic actions (those learned, routine

actions that are performed in context using actual tools and objects) bypasses attention

and accesses the output praxicon directly because the combined effects of the natural

context, tools, and objects may be enough information to activate that part of the motor

system specialized for overlearned, automatic movements (Paillard, 1982, Marsden,

1982, Rapscak et al., 1993, Schwartz & Buxbaum, 1997) The routes for routine and

non-routine actions both access the output praxicon but non-routine actions require

greater attentional resources than routine actions for accurate production. Additionally,

the route for novel actions accesses the innervatory patterns rather than the output

praxicon because by definition, "novel" movements have not been produced

previously and, therefore, would not have engrams represented in the output praxicon.

Novel actions also require more attentional resources than either routine or non-routine

actions.










LOW LEVEL SCHEMAS


HIGH LEVEL SCHEMAS


automatic

A \ routine
l-~A-



Outp t.r o ,. i \ non-routine.,
PraSicon m


^ f________\ novel __
Inn rvatory Patterns L -




|Motor Systems |


Figure 1-11. Breakdown of High Level Schemas.


Action

Schemas








Because the output praxicon receives information from the semantic system and the

semantic system is accessed by the perceptual systems, both systems (semantic and

perceptual) have been added to the model according to Rothi et al., (1991, 1997) in

Figure 1-12 and 1-13. Additionally, it is proposed that the executive functions in the

high level schematic system also communicates with the semantic system in order to

establish semantic relationships and interactions among related schemas. To

accommodate this communication, the semantic system has been divided into two parts

with one part connecting to the output praxicon and another part with connections to

and from executive functions. This division is not without debate, however, because

one could argue that the semantic system which communicates with the low level

movement schemas contains procedural information about tools and objects while the

semantic system which communicates with the high level action schemas contains

episodic information about action sequences.

Schwartz and Buxbaum (1997) and Reason (1979, 1984) also discussed visual,

auditory, and tactile perceptual systems as important input systems for online

monitoring and feedback to ensure correct action production. To accommodate these

reports, the perceptual system has been added to the high level action schema system

with input connections to executive functions (the planning and organization of action

schemas) and attention (Supervisory attention/working memory) in Figure 1-14. The

perceptual systems, therefore, send information to both the low (movement) and high

(action) schema systems.

In summary, the proposed relationship between apraxia and pragmatic action

described above separated the components of pragmatic action into low and high level













LOW LEVEL SCHEMAS I


Audltory/Verbail VbmalObject Ipus tul
Input Input Ip


HIGH LEVEL SCHEMAS


Figure 1-12. Addition of Perceptual Input to Low Level Schemas (Rothi et al., 1991, 1997).















LOW LEVEL SCHEMAS


Visual/Object
Input


I HIGH LEVEL SCHEMAS


VIsual/Gestund
Input


automatic


Figure 1-13. Addition of Semantic System to High Level Schemas.


Auditory/Verbal
Input















LOW LEVEL SCHEMAS I HIGH LEVEL SCHEMAS


Visual/Object Visual/Gstural
Input Input


Perceptual
Sytems
| V[~ ~ ~ y i V isu al A nally sis- ._______



SObjet --------oSyste
| Object RecognitionSyslem I -
Iput Ijroad"n
Frazicon ITA0
Ift 4 |-1 --- / A I et


Figure 1-14. Proposed Relationship Between Apraxia and Pragmatic Action.


Figure 1-14. Proposed Relationship Between Apraxia and Pragmatic Action.


AuditoryNVerbal
Input








schematic systems. The low level system targets planning, organization, production,

and comprehension of movement and movement sequences which are the final

common denominator of actions. The high level system includes planning,

organization, production, and comprehension of actions and action sequences

schemass) which originate in the complex schematic system of executive functions and

require attentional monitoring for correct production. The final common denominator

for both actions and movements is the praxis system, specifically, the output praxicon,

innervatory patterns, and motor system.

Predictions of the Proposed Relationship Between Movements and Actions

Based on the above description of the relationship between action schemas and

movements, the following predictions are made. First, pragmatic action can be divided

into two levels, a high level for actions and a low level for movements, with the final

common denominator of production through the output praxicon. Following this

prediction, there should be dissociations between the ability to plan and organize

actions (which occurs at the high level) from the ability to plan and organize

movements (at the low level). A deficit in the ability to plan, organize, and execute

actions has been documented by Mayer et al. (1990), Schwartz et al. (1991, 1993,

1995), and Schwartz and Buxbaum (1997) who studied patients with generalized brain

damage from closed head injury and labeled the deficit frontal apraxia, and Sirigu,

Zalla, Pillon, Grafman, Agid, and Dubois (1995, 1996), and Sirigu, Zalla, Pillon,

Grafman, Dubois, and Agid (1996) who studied patients with frontal lobe damage and

labeled the deficits as problems in managerial knowledge. However, Mayer et al.

(1990), Schwartz et al. (1991, 1993, 1995), Schwartz and Buxbaum (1997) and Sirigu








et al. (1995, 1996, in press) did not test for apraxia, a deficit in the ability to plan,

organize and execute movements. Ideomotor apraxia is well documented in the

literature by Liepmann (1900/1977, 1905b/1980b, 1907/1980c, 1920), DeRenzi (1985),

DeRenzi et al. (1966, 1982), Heilman et aL. (1982), Rothi et al. (1985), Rothi and

Heilman, (1985), Sirigu, Cohen, Duhamel, Pillon, Dubois, and Agid (1995), and

Foundas et al. (1995). Ideomotor apraxia has been localized to damage in the dominant

hemisphere (Liepmann & Maas, 1907; Geschwind, 1965; Heilman, 1979; and Poizner,

Meriens, Clark, Macauley, Rothi, & Heilman, 1997) and specifically to posterior

parietal cortex (Heilman, 1979; Freund, 1991; and Poizner et aL., 1997). However, the

focus of the above studies was on either high level action planning and organization

abilities, without documenting presence or absence of low level factors, or vice versa--

focus on low level action planning and organization abilities without documenting

presence or absence of high level factors.

Only one study to date has examined both low and high level schematic

functions within the same patient. Schwartz et al. (1995) examined action planning and

praxis in a patient, J.K., who suffered brain damage as the result of closed head injury.

Results indicated that J.K. evinced significant deficits in both the planning and

organization of actions and the planning and organization of movements.

Unfortunately, due to the extent of brain damage (lesions in right frontal lobe, bilateral

temporo-parietal lobe, and left occipital lobe) brain-behavior relationships could not be

made. Further studies are needed to document whether patients with impaired

planning and organization of actions have spared planning and organization of

movements and vise versa. However, because the supplementary motor area (SMA)








located in the prefrontal gyrus has been implicated in the production of learned

movements (Heilman, 1979) and it is proposed that the production of movements is the

final common pathway for all actions, researchers must be strict as to location of lesion

(i.e., with/without damage to SMA in patients with frontal lobe lesions) and in modality

of task (i.e., test action planning and organization through pictures and not through

actual production for patients with ideomotor apraxia as the ideomotor apraxia will

affect production) when selecting subjects and experimental tasks. It should also be

noted that due to the proposal that the output praxicon is the final common pathway for

production of actions, patients with ideomotor apraxia have the potential to be impaired

in production across tasks clinical and environmental. This impairment across tasks

has been documented experimentally by Foundas et al. (1995) and through case reports

by Jacobs, Macauley, Adair, Gold, Rothi, and Heilman (1995) and Sirigu et al. (1995).

A second prediction states that deficits in action semantics should affect the

functional use of tools within higher level schemas (in context or part of a goal-directed

schema) as well as within lower level schemas (out of context or independent of goal-

directed schema). Ochipa et al. (1989) reported a left-handed patient with a right

hemisphere lesion who demonstrated deficits in the functional use of tools across high

and low level tasks. High level tasks included eating a meal and low level tasks

included pantomiming to command (e.g., show me how you use scissors). Results

indicated that the patient was impaired in all tasks in which the functional relationship

between tool and object were important. It should be noted, however, that Ochipa et al.

(1989) labeled the patient's disorder ideationall apraxia" and defined it as a loss of

conceptual knowledge related to tool use. Raymer (1992) and Raymer and Ochipa








(1997), labeled this disorder "conceptual apraxia" which better describes this global

impairment of both low and high level schema functions.

A third prediction states that errors of action produced by non brain-damaged

people should relate to high level schema problems, the planning and organization of

actions, and not low level schema problems, the planning and organization of

movements. "Normal" people have intact praxis systems and should therefore produce

skilled movements accurately. However, the correct production of actions depends

upon the mediation of supervisory attention/working memory and if a person is not

"paying enough attention" or becomes distracted during the task at hand, a different

action, correctly produced, may be performed. Reason (1979, 1984, 1990) and Reason

and Mycielska (1982) studied human error and lapses of attention in everyday life.

They report that normal people produce errors of omission, addition, anticipation,

perseveration, capture, exchange, and substitution all of which occur within action

schemas and not within movements. For example, an error of capture occurred when a

lady went to the bathroom intending to brush her hair but she picked up the toothbrush

and began to brush her teeth. This lady did not pick up the toothbrush and brush her

hair or perform any action out of sequence for that particular schema, she merely

performed the wrong action schema for her intention which is a high level schematic

system error. Schwartz and Buxbaum (1997) classified the errors produced by brain

damaged patients during six daily activities, including making toast and wrapping a

present, using Reason's (1979, 1984) classification system. Schwartz and Buxbaum

(1997) reported that the brain damaged patients produced errors not only from the

categories described by Reason (1979, 1984) but also in four categories not described








by Reason (1979, 1984) grasp/spatial misorientation, spatial misestimation, tool

omission, and quality of movement. These four error types are hypothetically related to

the low level schematic system and would be predicted in patients with damage to the

praxis system by this model. Unfortunately, Schwartz and Buxbaum's (1997) patients

had heterogeneous etiologies and were not tested for conceptual or ideomotor apraxia

and therefore, direct evidence to support this prediction can not be obtained. Further

studies are necessary to examine the types of errors produced by brain damaged

patients and to establish a classification system using the framework of the proposed

model.

Classification of Action Errors

The types of errors described by Foundas et al. (1995), Ochipa et al. (1989),

Mayer et aL.(1990), Schwartz etal. (1991, 1993, 1995), Sirigu et al. (1995, 1996) and

Schwartz and Buxbaum (1997) can be divided into two categories based upon the low

(movement) and high (action) schema systems proposed by the relationship described

above (Figure 1-14). That is, error types that reflect deficits in the praxis production

systems (Rothi et al., 1991) would make up low level schematic system errors while

error types that reflect deficits in the planning and organization of action in the

executive function and attentional control mechanisms (Schwartz & Buxbaum, 1997)

would make up high level schematic errors. However, some types of errors could be

produced by both a high and a low level schema deficit. For example, the misuse of a

fork to stir tea could be explained by a deficit in the praxis conceptual system a low

level schema error or by a deficit in attentional control of the action sequences (the

patient had just finished using a fork to eat meat and did not switch tools before stirring








tea) a high level schema error. Other types of errors such as the body part as tool

errors reported by Rothi et al. (1991) can only be explained by a low level schema

deficit.

However, no study to date has measured deficits associated with low and high

level schemas within the same group of patients in order to tease out which action

errors are associated with which type of deficit. Foundas et al. (1995) and Ochipa et al.

(1989) tested the praxis production and conceptual systems (low level schema) while

Mayer et al., (1990), Schwartz et al., (1991, 1993), Sirigu et al. (1995, 1997), and

Schwartz and Buxbaum (1997) tested executive function, managerial knowledge, and

planning/ organizational abilities (higher level schema). Schwartz et al., (1995)

measured both low and high level schema abilities in a patient with diffuse brain

damage after closed head injury. However, an MRI of this patient showed focal lesions

in the left occipito-parietal area, right frontal area, and bilateral temporal areas which

negated any attempt to establish brain-behavior relationships and examine the

underlying nature of action errors.

Foundas et al. (1995) proposed that the errors produced by the experimental

subjects in their study could be divided into two groups production and conceptual.

Production errors consisted of movement, timing, and sequence errors and conceptual

errors consisted of misuse and mis-selection of a tool for the intended task. However,

as discussed above, production and conceptual errors can occur as the result of deficits

in low and/or high level schema systems. Production errors have been associated with

ideomotor apraxia resulting from discrete lesions of the dominant left hemisphere

(Foundas et al., 1995) as well as executive function disorders from closed head injury








(Schwartz et al., 1995) and conceptual errors have been associated with ideational/

conceptual apraxia resulting from discrete lesions of the dominant hemisphere (Ochipa

et al., 1989) as well as resulting from generalized brain damage from closed head injury

(Schwartz et aL., 1995). However, all of the experimental subjects in the Foundas et al.,

(1995) study had unilateral (dominant hemisphere) brain lesions resulting from single

strokes. Even though it would be easy to propose that the errors produced by these

subjects originated in the low level schema system, these subjects also evinced deficits

in the overall organization of meal planning, a high level schema system. However,

Foundas et al. (1995) did not test for presence of conceptual apraxia or executive

function (planning/ organizational) deficits that may explain the presence of high level

schema errors. Therefore, although Foundas et al. (1995) argued that the errors

produced by the experimental subjects strongly correlated with ideomotor apraxia, it is

not known whether conceptual or executive disorders also influenced the production of

action errors within these subjects.

Each of the above studies (Foundas et aL., 1995; Ochipa et al., 1989; Mayer et

aL, 1990; Schwartz et al., 1991, 1993, 1995; Sirigu et al. 1995, 1996, in press; and

Schwartz & Buxbaum, 1997) documents the presence of action errors in activities of

daily living. However, each study evaluated subjects with different etiologies of brain

damage (i.e., traumatic brain injury versus stroke) using different evaluation criteria

and asking different research hypotheses as to the nature of these action errors. As a

result, the nature of the movement and action errors produced by brain damaged

patients observed in natural contexts remains unclear. It is possible that by examining

deficits associated with both low and high level schemas in a homogeneous group of








subjects within a well controlled study that the nature of action deficits in activities of

daily living can be better understood and explained.

Statement of the Problem

While the presence of action and movement errors in natural contexts has been

described by several research groups, the mechanism of these action errors has not been

determined. In the above discussion, four explanations that may account for these

errors in natural contexts were described. The first proposes that movement and action

errors in natural contexts may be the result of an executive disorder of planning and

organization (Mayer et al., 1990; Schwartz et at., 1991, 1993, 1995; Sirigu et al., 1995,

1996; and Schwartz & Buxbaum, 1997), a high level schematic system function

according to the proposed relationship between movements and actions. The second

proposal states that action errors in natural contexts may be the result of an impaired

supervisory attention/ working memory (Schwartz et al., 1995; Schwartz & Buxbaum,

1997) a high level schematic system function according to the proposed relationship

between movements and actions. The third proposal states that action errors in natural

contexts may be the result of a deficit in action semantics or tool-function-object

knowledge (Ochipa et al., 1989, 1992) a function that implicates both high and low

level schematic system function according to the proposed relationship between

movements and actions. The fourth proposal states that the action errors in natural

contexts may be the result of ideomotor apraxia, an impairment in the production of

learned, skilled movements (Foundas et al., 1995); a low level schematic system

function according to the proposed relationship between movements and actions.








Previous studies of action errors compared the performance of left brain

damaged patients with normal controls (Foundas et al., 1995), compared patients with

heterogeneous etiologies with normal controls (Sirigu et al. 1995, 1996; Schwartz &

Buxbaum, 1997), or were single case studies (Ochipa et al., 1989; Mayer et al., 1990;

Schwartz et al., 1991, 1993, 1995). In fact, only three subjects, presented as single case

studies, form the basis of the Mayer et al., 1990, and Schwartz et al., 1991, 1993, and

1995 studies. The proposed cognitive neuropsychological model of pragmatic action

predicts that action errors can occur as the result of low level schematic system

impairment as well as high level schematic system impairment, both of which can be

associated with unilateral lesions of the dominant hemisphere. However, no study to

date has examined the presence of action errors in patients with unilateral lesions of the

nondominant (right) hemisphere in right-handed individuals. Due to the heterogeneous

nature of brain damage in previous studies as well as the lack of testing in both the low

and high level schema system, conclusions about brain-behavior relationships and the

nature of action errors can not be made. Therefore, studies are needed that 1) include

subjects with unilateral lesions of the nondominant (right) hemisphere and examine

both low and high level schema function; and 2) use a consistent error classification

system to score the subjects' performance in natural contexts which incorporates

scoring systems described thus far in the literature by Reason (1979, 1984), Foundas

et al. (1995), Schwartz et al. (1995), and Schwartz & Buxbaum, (1997); and 3)

compare subjects with brain damage limited to one hemisphere (right or left) from a

single event rather than generalized brain damage resulting from trauma in order to

examine brain-behavior relationships more clearly.








The purpose of this study, therefore, is to examine the nature (low versus high

level schema system functions) of action errors produced during an activity of daily

living in the natural environment in patients with left or right hemisphere brain damage

as well as neurologically normal control subjects. The following specific questions will

be addressed:

1. Does brain damage in either hemisphere result in production of
action errors in the natural environment or are action errors specific to left
hemisphere brain damaged patients?

The Foundas et al. (1995) study documented production of action errors by ten

left hemisphere damaged patients in the natural context of eating a meal. Other studies

described action errors produced by single subjects with either right hemisphere

damage (in a left-handed patient) or closed head injury (Ochipa et al., 1989; Mayer et

al., 1990; Schwartz et al., 1991, 1993, 1995; and Schwartz & Buxbaum, 1997).

Because only the left hemisphere damaged patients were experimentally studied, it is

not known whether the action errors described in the above studies are specific to

patients with lesions of the dominant hemisphere or can be found in right hemisphere

damaged patients as well. Using the proposed relationship between apraxia and

pragmatic action, it is proposed that patients with dominant (left) hemisphere brain

damage produce action errors consistent with both low and high level schema system

functions while patients with lesions of the nondominant (right) hemisphere produce

action errors consistent with high level schema system function only. This difference is

due to the bilateral representation of executive functions and the lateralized

representation of the visuokinesthetic engrams for learned movement to the left

hemisphere. This difference would predict that patients with left hemisphere damage








have more opportunity for errors and may therefore produce more errors than patients

with right hemisphere lesions. According to the proposed model, patients with lesions

of the nondominant right hemisphere may evince action errors as a result of executive

system function impairment (deficits in the planning, and organization of actions) as

well as from attention system impairment as these are both represented bilaterally and

impairment of one hemisphere may result in deficits in these systems. The errors

resulting from attentional system impairment may include spatial errors resulting from

neglect which has been described as a disorder of attention by Brain (1941), Zangwill

(1944), McFie, Piercy, and Zangwill (1950), Denny-Brown and Banker (1954),

Heilman (1979), and Heilman, Watson, and Valenstein (1985). Heilman (1979)

reported that some patients with neglect fail to eat from one side of their plate but have

spared action plans for eating and use of tools during eating. The null hypothesis of no

difference in number of action errors produced while eating a meal between the groups

(left hemisphere damaged subjects, right hemisphere damaged subjects, and controls)

will be tested.

2a. Does presence of deficits in
1. production of learned skilled movements,
2. conceptual knowledge of tool-function relationships,
3. action planning and organization, and
4. supervisory attention/working memory
predict production of action errors in the natural environment?

2b. And if so, to what extent does each type of deficit (low level system
deficits--1 and 2 above and high level system deficits-3 and 4 above) predict
type of error (including misuse, mis-selection, external configuration, internal
configuration, timing, quantity, sequence [omission, addition, sequence],
exchange, substitution, movement, and body part as tool)?

Deficits in conceptual knowledge of tool-function relationships and deficits in

production of learned skilled movements are proposed to result from lateralized lesions








to the dominant (left) hemisphere in most right-handed people (Rothi & Heilman, 1985)

and are hypothesized to be associated with low level schemas in the proposed

relationship between movements and actions. Deficits in the planning and organization

of actions and deficits in supervisory attention/working memory are proposed to result

from left, right, and bilateral lesions as well as degenerative brain damage (Schwartz et

al., 1991, 1993, 1995) and are hypothesized to be associated with high level schemas in

the proposed cognitive neuropsychological model of pragmatic action. As a result, all

four deficits of higher cortical function described above have the potential to produce

action errors in natural contexts. To examine the relationship between these disorders

of higher cortical function and production of action errors in the environment,

independent measures (non-context dependent) of each disorder will be correlated with

total number of action errors produced in the experimental measures (context-

dependent). The null hypothesis of no significant correlation between any one

independent measure of higher cortical function and number of total action errors

produced during the context dependent activities of daily living will be tested.

In addition, it is hypothesized that tool and nontool errors will be produced to

different degrees by patients depending on the type and severity of higher cortical

function deficit. That is, patients with deficits in the praxis production and conceptual

systems (low level movement schema) may produce more tool errors (misuse, mis-

selection, internal configuration, external configuration, quantity of food on a utensil,

movement associate with the tool, and body part as tool) than patients without a deficit

in the praxis production and/or conceptual system. Hypothetically, the number of tool

errors produced should reflect the severity of the underlying deficitss. Patients with a








deficit in the planning and organization of actions and/or supervisory attention/working

memory (high level schema) may produce more nontool errors (exchange, substitution,

sequence, omission, addition, and timing,) than patients without a deficit in the

planning and organization of actions and/or supervisory attention/working memory.

Although these patients may also show tool errors of misuse and mis-selection which

could arise from high level schema deficits, these patients should not evince tool errors

of internal configuration, external configuration, quantity, and body part as tool without

a concomitant deficit in the praxis system.

As a result of the action error classification system described above, it is

proposed that the errors reported by Foundas et aL. (1995) as resulting from ideomotor

apraxia or production disorders, may also have been influenced to a certain extent by

high level schema deficits and that errors reported by Mayer et al. (1990), Sirigu et al.

(1995a, 1995b, 1996), and Schwartz et al. (1991, 1993, 1995) as resulting from

disorders of action planning and organization and/or supervisory attention/working

memory may also have been influenced by production and conceptual disorders.

However, it is not known which of the four disorders (praxis production, praxis

conceptual, action planning/ organization, and supervisory attention/working memory)

is most influential in inducing action errors in natural contexts or if all four disorders

are equally influential. To answer this question, independent measures of the praxis

production system, praxis conceptual system, action planning/organization, and

supervisory attention/working memory will be correlated with frequency of each type

and classification of error produced during the experimental activities of daily living.





62


The null hypothesis of no significant correlation between any one independent measure

and number of action errors within one type/classification will be tested.













CHAPTER TWO
METHODS


Some patients with brain damage produce action errors during activities of daily

living (Mayer et aL., 1990; Schwartz et aL., 1991, 1993, 1995; Schwartz & Buxbaum,

1997; Ochipa et al., 1989; Foundas et aL., 1995). Investigators in recent research reports

have examined the planning, organization, and production of these action errors through

different theoretical models with different inherent assumptions and predictions, using

different research hypotheses and evaluation criteria, and included subjects with

heterogeneous etiologies of brain damage (Mayer et al., 1990; Schwartz et al., 1991,

1993, 1995; Schwartz & Buxbaum, 1997; Sirigu etal., 1995, 1995, 1996, in press;

Ochipa et al., 1989; Foundas et al., 1995). The current study describes a proposed

relationship between apraxia and pragmatic action which serves as a framework for the

examination of the nature of movement and action errors produced by patients with

unilateral left or right hemisphere brain damage as well as matched normal control

subjects.

Subjects

Three subject groups (two experimental groups and one control group)

participated in this study. The first experimental group consisted of patients with left

hemisphere strokes (LBD) and the second experimental group consisted of patients with

right hemisphere strokes (RBD). All experimental subjects were at least one month post

onset of their stroke, were right-handed, and native English speakers. None had a history
63








64

of previous neurological disease. The experimental subjects were at least one month post

onset of a unilateral CVA (from a single event) as documented by CT/MRI scan and were

identified and recruited through the following sources: Veterans Affairs Medical Center,

W. A. Shands Teaching Hospital, and UpReach Rehabilitation Center, Gainesville,

Florida; Sacred Heart Medical Center, Deaconess Medical Center, St. Luke's

Rehabilitation Center, and University Programs in Communication Disorders Speech and

Hearing Clinic, Spokane, Washington. CT/MRI studies done at least two weeks post-

stroke were obtained for each experimental subject and reviewed by a neurologist. The

neurologist classified the lesions as either right or left as defined by relationship of lesion

to longitudinal fissure. The lesion locations were used to classify the experimental

subjects into left-hemisphere or right-hemisphere damaged groups.

The LBD experimental group included 20 subjects whose ages ranged from 32 to

79 (M = 60.2; SD = 13.50). Educational level ranged from 3 to 16 years (M = 11.85; SD

= 3.25). Time post stroke varied from one month to 261 months (M = 41.45; SD =

65.28). Descriptive information for these LBD subjects is provided in Table 2-1.

The RBD experimental group included 10 subjects whose ages ranged from 54 to

77 (M = 64.2; SD = 8.11). Educational level ranged from 8 to 16 years (M = 12.60; SD =

2.46). Time post stroke varied from one month to 40 months (M = 11.3; SD = 15.47).

Descriptive information for these RBD subjects is provided in Table 2-2.













Table 2-1 LBD Subject Identification


Subject Sex Age Education MPO*


LI
L2
L3
L4
L5
L6
L7
L8
L9
LIO
Lll
L12
L13
L14
L15
L16
L17
L18
L19
L20


28
53
25
84
2
168
67
261
19
16
37
28
1
4
17
1
1
1
8
8


*MPO months post-onset













Table 2-2 RBD Subject Identification


Subject Sex Age Education MPO*


RI M 76 13 40
R2 M 58 10 2
R3 M 59 16 27
R4 M 77 13 5
R5 M 62 12 4
R6 M 73 16 1
R7 F 54 8 1
R8 F 61 14 1
R9 M 62 12 1
RIO M 60 12 31
*MPO months post-onset


The control subject group consisted of 10 neurologically intact community

volunteers from Gainesville and Bradenton, Florida, and Spokane, Washington. All

control subjects were also right-handed and native English speakers. Their ages ranged

from 46 to 76 years (M = 64.1; SD = 8.62). Educational levels ranged from 10 to 16

years (M = 12.6; SD = 1.84). Descriptive information regarding the control subjects is

provided in Table 2-3.















Table 2-3 Control Subject Identification


Subject Sex Age Education


Cl M 69 13
C2 M 71 13
C3 M 63 12
C4 M 59 16
C5 F 69 12
C6 F 63 10
C7 M 68 14
C8 M 57 14
C9 M 76 10
C10O F 46 12


Materials and Methods

The description of the experimental procedures that follows is divided into two

sections, each containing a description of the materials, methods of presentation, and

scoring procedures relevant to each measure. The first section, entitled evaluation of

independent variables describes tests designed to measure the deficits associated with

inpairment in action planning and organization (i. e., "frontal apraxia," a high level

schema deficit), deficits in the production of learned, skilled movements (i.e., ideomotor

apraxia, a low level schema deficit), deficits in conceptual knowledge of tool-object-

function relationships (i.e., conceptual apraxia a low and/or high level schema deficit),

and deficits in supervisory attention/working memory (high level schema deficit). In








68

addition, tests of language abilities and presence of neglect are described. Results of the

tests of language and neglect were not used to directly answer the stated research

questions. Rather, due to the co-occurrence of aphasia and apraxia in left hemisphere

damaged patients and the occurrence of neglect in right hemisphere damaged patients,

these variables must be systematically measured in order to assist in the interpretation of

the results. The second section, entitled evaluation of dependent variables describes a

task designed to measure production of movement and action errors in an activity of daily

living.

All subjects were administered the same battery of tests. Testing occurred within

one day or across a two-day period. The order of presentation of all tests was

randomized across subjects.

Evaluation of Independent Variables.

Disorders of action planning and organization

Materials. The Tower of London task (TOL, Shallice, 1982; Krikorian, Bartok, &

Gay, 1994) was given to evaluate deficits in the ability to plan and organize actions

associated with "frontal apraxia." The TOL consists of three pegs of different lengths

mounted on a strip of board and three colored balls (red, green, and blue). The balls have

holes drilled through them so they can be placed on the pegs. The left peg held all three

balls, the middle peg held two balls, and the right peg held one ball.

Methods. The standardized administration procedures described by Krikorian et

al. (1994) were used in this study. The balls were placed in a standard initial position

(red ball on top of green ball on left peg and blue ball on middle peg) and the subject was








69

asked to manipulate the balls on the pegs to reproduce a pictured end state (Figure 2-1).

The subject performed twelve trials of increasing difficulty. A problem was correctly

solved when the end state was achieved in the prescribed number of moves. The subject

had three attempts to solve each trial.

Scoring. The standardized scoring procedures described by Krikorian et al.

(1994) were used in this study. The subject received three points if the trial was solved

correctly on the first attempt; two points if solved on the second attempt; one point if

solved on the third attempt; and zero points if not solved by the third attempt. The

examiner scored each subjects' performance online according to the above mentioned

criteria and percent correct was calculated by dividing the number of points accumulated

by the total number of points possible.

Disorders of learned, skilled movements

Materials. The Randomized Form A (RFA) and Randomized Form B (RFB)

versions of the Gesture -to Command subtest of the Florida Apraxia Battery: Experimental

Edition (Appendices A & B, Rothi, Raymer, Ochipa, Maher, Greenwald, & Heilman,

1992) were given to evaluate presence and severity of disorders of learned, skilled

movements associated with ideomotor apraxia. The Gesture To Command subtest

consists of 30 commands which elicit transitive or intransitive gestures. Transitive

gestures (e.g., show me how you use a hammer) involve tool knowledge and

demonstration of how one uses tools. Intransitive gestures (e.g., show me how you

would salute) do not involve tool knowledge and have been defined as emblems by

Lemay, David, and Thomas (1989). (See Appendix A for copies of the RFA and RFB).















R

B G


Initital State






G

B R



End State 3 moves









End State 5 moves

End State 5 moves


Figure 2-1. The Tower of London (from Krikorian, Bartok, & Gay, 1994).












To control for the possible effects of hand used to gesture, Raymer, Maher,

Macauley, Foundas, Rothi, and Heilman (1997) studied the performance of 16

neurologically normal, right handed control subjects on this Gesture to Command subtest

looking at the influence of hand used. They divided the 16 subjects into four groups and

administered the RFA or RFB counterbalancing hand used to gesture in the following

manner: Group 1 was administered RFA and gestured all items with the right hand

followed by all items with the left hand; Group 2 was administered RFA with the left

hand used first; Group 3 was administered RFB with the right hand used first; and Group

4 was administered RFB with the left hand used first. The tapes were then scored

pass/fail by two raters with at least two year's experience in scoring gesture production

using the Rothi, Mack, Verfaellie, and Brown (1988) error pattern analysis. Results

indicated that there was no significant difference in the performance of the right and left

hands by control subjects with no history of neurological disease when producing

gestures to command (Raymer et al., 1997).

Methods. Each subject was administered either the RFA or RFB using the

instructions written at the top of each score sheet (See Appendix A). The RFA and RFB

were alternated across subjects with half receiving RFA and half receiving RFB. To

control for possible effects of hemiparesis, the experimental subjects produced the

gestures using their ipsilesional hand and five of the ten control subjects were asked to

gesture using their non-dominant left hand. The subjects were videotaped while

performing these gestures.








72

Scoring. The videotapes were scored by two examiners familiar with ideomotor

apraxia who underwent a training session similar to that described by Maher (1995).

During the training sessions, the examiners were required to obtain at least 85%

agreement with practice tapes before beginning to score the experimental tapes. The

scoring system followed the error pattern analysis described by Rothi et al. (1988). Each

gesture was given a pass/fail rating and percent correct was calculated.

Inter-rater reliability was calculated by having the raters score four videotapes

twice and calculating percent agreement across the two sessions. Intra-rater reliability

was calculated by calculating percent agreement between the two scorers on 20% of the

videotapes.

Disorders of tool-object-function knowledge

Materials. A revised version of the Tool-Object Matching Test (TOM) described

by Ochipa et al. (1989) was given to evaluate deficits in tool-function-object knowledge

associated with ideational/conceptual apraxia (Appendix B). The original version of the

Tool-Object Matching Task described by Ochipa et al. (1989) used randomly generated

tools as foils. The revised version used the same foil tools for each trial across subjects.

The foil tools were drawn from the possible correct tool choices for the other trials and all

tools were used three times once as a correct answer and twice as foils (except for

hammer which is a correct answer twice and used as a foil four times).

Methods. Each subject was presented with a partially completed task (e.g. a

partially sawed board) and an array of three tools consisting of the target tool and two foil

tools (e.g. a hammer, saw, and pencil). The subject was asked to select the appropriate








73

tool to complete the task according to the instructions written at the top of the score sheet

(Appendix B).

Scoring. The examiner scored the subject's responses online using a binomial,

pass/fail criteria and percent correct responses was calculated.

Disorders of supervisory attention/working memory

Materials. The Trailmaking Test Parts A & B (Reitan, 1944) was used to

evaluate deficits in supervisory attention/working memory. The Trailmaking Tests -

Parts A & B ("Trails A & B") are standardized neuropsychological measures that are

sensitive to brain damage. Good performance on Trails A & B depends upon attention to

task and working memory with strong visual search and motor performance components.

The standardized test sheets were used in this study. Trails A & B each consist of a

single sheet of 8 1/2 x 11 white paper with the numbers 1 25 (Part A) and 25 numbers

and letters (Part B) written in random order on the page. The numbers and letters are

written in circles.

Methods. Trails A & B were administered according to the standardized

instructions by the examiner.

Scoring. Trails A & B were scored according to the standardized instructions by

the examiner. The time in seconds required for the subject to complete Part A and Part B

were recorded separately and compared to normative data reported by Davies (1968)

(Appendix C).










Disorders of language

Materials. The Western Aphasia Battery (WAB, Kertesz, 1981) was used to

evaluate presence, type, and severity of aphasia. The WAB is a standardized test of

acquired language disorders which enables the calculation of aphasia severity through an

aphasia quotient (AQ) and determination of aphasia type through the WAB aphasia type

taxonomy. The WAB is divided into four sections, spontaneous speech, comprehension,

repetition, and naming, which test each aspect of verbal language skills in brain-injured

patients.

Methods. The WAB was administered according to the standardized instructions

by the examiner, a certified speech-language pathologist with seven years experience

working with neurologically based communication disorders.

Scoring. The WAB was scored according to the standardized instructions by the

examiner, a certified speech-language pathologist with seven years experience working

with neurologically based communication disorders. An aphasia quotient reflecting

aphasia severity score was calculated and type of aphasia was documented based upon

the WAB aphasia taxonomy (Kertesz, 1981).

Neglect

Materials. A line bisection task was used to evaluate presence and severity of

neglect. For the current study, three lines measuring eight, ten, and twelve inches in

length were drawn horizontally across the middle of individual 1 1x14 pieces of paper

using a thick black marker.








75

Methods. The examiner placed one of the pieces of "lined" paper described

above horizontally on the table approximately eighteen inches in front of the subject with

the middle of the line even with the subject's midsagittal plane. The subject was given a

pen and asked to "mark the center of the line." This procedure was repeated for each of

the three lines. The lines were given in random order to each subject

Scoring. The examiner scored the accuracy of the subject's mark by measuring

the difference between the true middle of the line and the subject's mark. An average

difference was calculated in centimeters. Presence of neglect was determined by a

difference score of 10mm between the true midline and the subject's mark (Heilman,

Watson, & Valenstein, 1985). Severity of neglect was determined by the magnitude of

the average difference in that the greater the difference, the more severe the neglect.

Evaluation of Dependent Variable-Eating a Meal.

Materials. A meal consisting of a main course, at least one side dish, a dessert,

and a beverage, and utensils such as a knife, fork, and spoon, were organized on a

hospital tray according to standard etiquette rules. The food was placed on standard

hospital plates with covers. Additionally, three foil items a toothbrush, comb, and

pencil, were also placed on the tray interspersed in random order with the utensils. The

location of the foil items and the utensils were ipsilateral to side of lesion for the

experimental subjects and counterbalanced across control subjects. Subjects who were

inpatients received their meals in their hospital rooms from food service personnel during

regular mealtimes. Subjects who were not inpatients obtained their meals through the

hospital cafeteria using similar plates and covers as the inpatients and ate privately in a








76

research laboratory. Straws, condiments, and napkins were made available to each

subject by the examiner.

Methods. All subjects were videotaped while eating the meal. The videotaping

was done using a Sharp videocamera set on a tripod approximately five feet in front of

the subject. Taping was done on a TDK 120 minute videotape. The video camera was

turned on just prior to the subject receiving their food and turned off after completion of

the meal. The examiner was not in the room during the videotaping and nurses, spouses

or significant others were instructed not to assist the subject in any way during the eating

of the meal, unless specifically asked to do so by the subject. Conversation with family

members during the meal was allowed to ensure a comfortable, natural environment.

The examiner arranged the utensils and foil items on the tray after the meal had been

delivered counterbalancing location (right or left) across control subjects. Subjects with

specific dietary restrictions were accommodated through choice of meal. In order to

control for possible effects of hemiparesis, an equal number of control subjects were

asked to eat with their nondominant hand as experimental subjects who ate with their

nondominant hand.

Scoring. The mealtime videotapes were scored by two raters and a trainer at the

same time but independently. The scoring was completed in a quiet room using a 20 inch

viewing monitor and a Sony video player with slow motion and freeze frame capabilities.

The trainer controlled the tape player for the raters but the raters were free to stop and

rewind to view an action or series of actions as many times as necessary to ensure that all

actions produced by the subjects were included. The raters were encouraged to consult a








77

list of error types given to them during the training time as often as needed during the

scoring periods. No discussion was allowed with respect to the subject's performance

and the raters were not able to see each other's or the trainer's score sheet.

The scoring system described by Foundas et al. (1995) and the error classification

system associated with the proposed relationship between apraxia and pragmatic action

described in chapter one was used in this study. For the current study, an action was

defined as a sequence of individual movements that result in the accomplishment of a

definable goal and movements were defined as changes in place, position, or posture.

For example, picking up the fork, piercing a piece of meat, and bringing the meat to the

mouth would be three individual movements that make up one action. Each action was

categorized as a tool or a nontool action. Examples of tool actions would be stirring tea

with a spoon; eating a piece of meat with a fork; and buttering a slice of bread with a

knife. Examples of nontool actions would be placing the napkin in the lap; moving an

empty plate to the side of the tray; and opening a packet of sugar. The raters were

instructed to use a standard recording form to record the quadrant of the tray in which the

action was initiated, which hand was used to perform the action, which tool and object

were used, and which action was performed. The raters then determined if the action was

correct or incorrect and ifjudged incorrect, the raters assigned an error type to the

incorrect action. The error types are described below:

Errors proposed to result from Low Level System Impairment:

internal configuration: hand posture used to manipulate
tool was incorrect (e.g., holding a spoon with a tightened fist).








78

external configuration: action produced was misplaced in space in absence of
perceptual deficits (e.g., scooping with a spoon on the table beside the plate).

body part as tool: action was produced using a body part as the tool (e.g.,
spreading butter on bread with index finger).

timing: incorrect timing during a sequence of movements within one action (e.g.
spreading butter on bread with pauses in between the knife movements) or
between two actions (e.g., holding a spoonful of food close to the mouth while
one is still chewing a previous bite).

quantity: incorrect or inefficient amount of food is taken to the mouth (e.g.,
putting too much or too little mashed potatoes on a spoon which is then
brought to the mouth).

movement: movement produced was inaccurate or incorrect (e.g., using wrist
motion rather than shoulder motion)

Errors proposed to result from high level system impairment:

sequence: sequence of actions was incorrect (e.g., stirring tea before
adding the sugar)

omission: omitting one action from a sequence of actions (e.g., putting
sugar in tea and not stirring it at all).

addition: adding an additional action during a sequence of actions (e.g.
putting sugar in tea and eating a bite of food before stirring the tea).

Errors proposed to result from high and/or low level system impairment:

misuse: action (movement) produced was incorrect for chosen tool (e.g., stirring
with a knife) (See Figure 1-7.)

mis-selection: action (movement) produced was correct for the chosen tool but
not for the goal with the object (e.g., using a fork to eat ice cream) (See Figure 1-
7.)


It was possible for one action to be judged as represented by more than one error type

during the scoring procedure.








79

Actions over the entire meal were recorded and scored and the following

calculations were obtained:

1. percent correct tool actions:

# of correct tool actions
total # of tool actions

2. percent correct nontool actions:

# of correct nontool actions
total # of nontool actions

3. percent correct total actions:

# of total correct actions (tool + nontool)
# of total actions (tool + nontool)

4. action per time ratio:
total time of meal (seconds)
total # of errors over the entire meal

Rater Training

The rater training protocol used in this study was based on the protocol described

by Maher (1995). The raters for this study were two graduate students in speech-

language pathology. The raters were novices in the study of apraxia and unfamiliar with

the specific research questions in the study. The raters were told that every action the

subjects produced was important and to make sure that every action was recorded and

scored. Prior to a scoring session, the raters were given a score sheet and a list of error

types with definitions and examples. They were asked to record and score a two minute

portion of a mealtime tape from a previous study. These recordings and scores were

reviewed with the raters by the examiner. Any discrepancies in scoring were reevaluated








80

by watching the pertinent section of the videotape and then discussing the scores. This

procedure continued until 85% or greater reliability between the raters was achieved over

16 trials (one trial = one action). At that point, the error types were reviewed by the

trainer and recording/scoring of test trials began.

Inter-rater reliability was calculated by selecting four videotapes at random and

calculating point-to-point reliability for the raters. The raters were not aware of which

subjects were used for reliability calculations. Two subjects selected at random were

scored twice by the raters to calculate intra-rater reliability.

Statistical Analyses

The statistical analyses were tailored specifically for each of the research

questions. For question number 1

Does brain damage in either hemisphere result in production of action errors in
the natural environment or are action errors specific to left hemisphere brain
damaged patients?

the LBD, RBD, and control subjects were compared using an analyses of variance with

the dependent variable being total number of action errors produced during the eating

task. Because a negatively skewed distribution was found for the control subjects, a

nonparametric test, the Kruskal-Wallace, was used. The Wilcoxon Rank Sums Test was

then used to determine differences between groups.

For question 2a
Does presence of deficits in
1. production of learned skilled movements,
2. conceptual knowledge of tool-function relationships,
3. action planning and organization, and
4. supervisory attention/working memory
predict production of action errors in the natural environment?








81

Pearson Product Moment Correlations were calculated between the independent

experimental measures (TOL, FAB, WAB, Trails A, Trails B, Neglect, Tool-Object

Matching) and total number of action errors produced during the meal.

For question 2b

And if so, to what extent does each type of deficit (low level system deficits--1
and 2 above and high level system deficits--3 and 4 above) predict type of error
(including misuse, mis-selection, external configuration, internal configuration,
timing, quantity, sequence [omission, addition, sequence], exchange, substitution,
movement, and body part as tool)?


A multiple stepwise regression statistic was calculated between the independent measures

(TOL, FAB, Trails A, Trails B, Tool-Object Matching) and the dependent variables of

types of errors produced during the meal (misuse, mis-selection, external configuration,

internal configuration, timing, sequence [omission, addition, sequence], movement, and

body part as tool.













CHAPTER THREE
RESULTS

The purpose of this study was to examine the nature of movement and action

errors produced by brain-damaged patients during an activity of daily living--eating a

meal. Two experimental subject groups and one control subject group participated in this

study. The first experimental group consisted of twenty left hemisphere brain-damaged

patients; the second experimental group consisted often right hemisphere brain-damaged

patients; and the control subject group consisted often neurologically normal control

subjects. There were no significant differences between the three groups in age F(2, 37)

= 0.61, p > 0.54 or educational level F(2, 37) = 0.36, p > 0.69. There was also no

significant different in the time post stroke between the two experimental groups F(1, 28)

=2.04,p> 0.16.

Experimental Results

Errors of movement and action were theorized to occur at two levels based on the

low (movement)'and high (action) level schema systems proposed by the relationship

described in Figure 1-4. That is, error types that reflect deficits in the praxis production

system (Rothi et al., 1991) would make up low level schematic system errors while error

types that reflect deficits in the planning and organization of action in the executive

function and control mechanisms (Schwartz & Buxbaum, 1997) would make up high

level schematic errors. As described in Chapter Two, the evaluation of independent

variables included tests that were proposed to measure the different deficits that can








occur within the two levels. The Tower of London (TOL; Shallice, 1982, Krikorian et

al., 1994) was given to evaluate deficits in the ability to plan and organize actions

associated with "frontal apraxia", a high level deficit. The Gesture to Command subtest

of The Florida Apraxia Battery: Experimental Edition (Rothi et al., 1992) was given to

evaluate the presence and severity of disorders of learned, skilled movement associated

with ideomotor apraxia, a low level deficit. A revised version of the Tool-Object

Matching Test (TOM) described by Ochipa et al., (1989) was given to evaluate deficits in

tool-function-object knowledge associated with ideational/conceptual apraxia, a low level

deficit. Additionally, the Trailmaking Test Parts A & B (Reitan, 1944) was given to

evaluate deficits in supervisory attention/working memory; The Western Aphasia Battery

(WAB; Kertesz, 1981) was given to evaluate presence of aphasia; and a line bisection

task was used to evaluate presence of neglect. Descriptive statistics for the three groups

on each of the above tasks are listed in Table 3.1. Overall, the three groups were

prototypic of their constituents. That is, 70% of the RBD had neglect while only 40% of

the LBD and none of the controls had neglect. 85% of the LBD had aphasia while none

of the RBD or controls were aphasic. 65% of the LBD had ideomotor apraxia while 20%

of the RBD and none of the controls were apraxic. 20% of the LBD and none of the

RBD or controls had conceptual apraxia. In addition, 80% of LBD and 70% of RBD had

"frontal apraxia" and were slower on the attention tasks than the controls.

Neglect and aphasia were ruled out as variables that affected the data by dividing

all subjects into two groups according to presence/absence of neglect and

presence/absence of aphasia and comparing the mean number and types of errors

produced by the two groups. No significant differences between the groups were found.








Table 3.1 Descriptive Statistics for Each Group on Measures of Independent Variables

Task Mean SD Max Min

Tower of London
LBD 24.4 11.21 34 3
RBD 24.1 10.21 33 5
Control 30.8 3.08 35 26
Total Possible 36

Gesture to Command
LBD 12.8 6.03 25 3
RBD 20.8 6.01 27 10
Control 23.0 3.13 28 17
Total Possible 30

Tool-Object Matching
LBD 9.3 1.94 10 3
RBD 10 0 10 10
Control 10 0 10 10
Total Possible 10

Trails A
LBD 86.71 56.98 232 26
RBD 112.2 74.68 231 39
Control 49.2 23.55 95 21
Range for NBD* varies according to age, (Appendix C)

Trails B
LBD 208.1 197.45 925 67
RBD 213.5 154.15 561 100
Control 111.3 47.78 199 66
Range for NBD* varies according to age, (Appendix C)

Western Aphasia Battery
LBD 69.54 27.20 98.2 10.8
RBD 97.92 1.14 99.6 96.1
Control 99.02 0.64 99.8 98.2
Total Possible 100

Line Bisection
LBD -1.6 1.56 13.3 -31.04
RBD 8.9 2.87 25.0 -2.33
Control 2.9 0.39 0.67 -7.00
Range for NBD*
*NBD= non brain-damaged individuals








Research Questions

Research Question #1

Does brain damage in either hemisphere result in production of action errors in
the natural environment or are action errors specific to left hemisphere brain
damaged patients?

To determine if patients with LBD (left brain damage) and RBD (right brain

damage) produce action errors in the natural environment, the following null hypothesis

was tested:

Ho: There is no significant difference between the mean number of errors
produced while eating a meal for the LBD, RBD, and control groups (NBD, non-
brain damaged).

Using a Kruskal-Wallis One-Way Analysis of Variance on Ranks (due to the

skewdness of the control group having produced no errors during the meal), the mean

number of total errors was compared between the three groups. A significant between

group difference was found (Table 3.2).



Table 3.2 Overall Kruskal-Wallis ANOVA for Errors Produced During the Meal

ANOVA dF (2.2)
Task Chi-Square H Statistic P-Value _

Total Errors 16.091 0.0003*

Tool Errors 10.496 0.0053*

Non-Tool Errors 15.479 0.0004*
* significant at p>0.01


Because a significant difference was found for total number of errors produced

during the meal, a follow-up analysis was conducted using the Kruskal-Wallis Multiple

Comparison Z-Value Test. This test is used to examine group differences with non-









parametric data. Results of the follow-up analyses, comparing each brain-damaged group

with the normal controls are listed in Table 3.3. For total number of errors produced

during the meal, the LBD and RBD groups did not differ from each other. However,

both groups did differ significantly from the control group.



Table 3.3 Comparison of LBD and RBD Groups with Normal Controls


Group I vs. Group 2

Total Errors
LBD vs. RBD
LBD vs. Controls
RBD vs. Controls

Tool Errors
LBD vs. RBD
LBD vs. Controls
RBD vs. Controls

Non-Tool Errors
LBD vs. RBD
LBD vs. Controls
RBD vs. Controls

*significant at p>0.01


Critical Difference


z-value >1.96
z-value >1.96
z-value >1.96


z-value >1.96
z-value >1.96
z-value >1.96


z-value >1.96
z-value >1.96
z-value >1.96


Actual Difference


1.012
3.997*
2.585*


2.269*
2.937*
0.5783


0.568
3.861*
2.851*


In order to determine if the significant difference found for number of total errors

could be accounted for by differences in number of tool errors and/or nontool errors

produced by the three groups, additional ANOVAs were calculated for total number of

tool errors and nontool errors. Significant group differences were found for both tool and

nontool errors. See Table 3.2 for ANOVA results. To examine between group

comparisons, follow-up analyses were conducted using the Kruskal-Wallis Multiple