A NEUROBEHAVIORAL INVESTIGATION OF EMOTION PROCESSING AND INTERPERSONAL FUNCTIONING IN ALCOHOLISM By LAUREN A. HOFFMAN 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 2017
2017 Lauren A. Hoffman
To Lucy (AKA Goose)
4 ACKNOWLEDGMENTS First, I would like to acknowledge my advisor, Dr. Sara Jo Nixon. Her guidance throughout this journey has been invaluable and I am certain that it will continue to be hereafter. Mentoring in the midst of maint aining a successful research career is no small endeavor. Yet, Dr. Nixon manages to do it with grace I am exceedingly grateful for her mentorship and I strongly believe that my success to date would not have been possible without it. Thanks to her tutelag Relatedly, gratitude is owed to past and present members of the Neurocognitive Laboratory, including Robert Prather, Layla Lincoln, Dr. Ben Lewis, Dr. Alfredo Sklar, Dr. Jeff Boissoneault, Julianne Price, Sar ah Reaves, Ian Frazier and Christian Garcia. for facilitating my understanding of electrophysiological research, Dr. Lewis, Mr. Prather and Mr. Frazier for assisting in task development, as well as Ms. Lincoln for her essential contributions to data collection. extend my gratitude to colleagues and friend s outside of the lab who selflessly offered their assistance throughout the dissertation process. Their support was tremendously appreciated. Furthermore, I would like to recognize my committee members, Drs. Jesse Dallery, Andreas Keil, and Catherine Striley for their guidance and encouragement. I committee and I am sincerely grateful for their unyielding support and confidence in my abilities Additionally, this research would not have been possible without the support of the University of Flori da Department of Psychiatry, the National Institute on Alcohol Abuse and Alcoholism (R01AA022456 01; PI Nixon), and the participating treatment facilities and clients.
5 Finally, I must extend my profound gratitude to family and close friends. Foremost, to my brilliant father whose support and encouragement ha ve been essential to my progress and accomplishments. As the saying goes from the tree and indeed, my educational and professional drive is entirely owed to him. to dedicate my life to addictions research. Her demonstration of strength and perseverance in recovery is admirable and serves as personal motivation to better understand substance use disorders and help others affected by them A special thank you goes out to my wonderful sister Jessica close friend Stephanie and significant other Timothy, who s e unwavering support and love has been a tremendous comfort. I could not have managed this journey without them Last but certainly not least, a most deserved th ank you is owed to my dogs, Lucy and Gunner, for being the best companions a dissertation writer could ask for. They are a constant reminder to me that this short life is best lived with love, light hea rted fun, and plenty of treats.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Alcohol Use Disorders Defined ................................ ................................ ............... 13 Neurobehavioral Correlates of Alcohol Use Disorders ................................ ............ 14 Emotion Processing and Alcohol Use Disorders ................................ ..................... 17 Summary ................................ ................................ ................................ ................ 35 2 STUDY AIMS AND METHODS ................................ ................................ .............. 37 Stud y Aims ................................ ................................ ................................ .............. 37 Aim 1 ................................ ................................ ................................ ................ 39 Aim 2 ................................ ................................ ................................ ................ 40 Summary of Aims ................................ ................................ ............................. 40 Methods ................................ ................................ ................................ .................. 41 Participants ................................ ................................ ................................ ....... 41 Laboratory Protocol ................................ ................................ .......................... 44 Self Report Instruments ................................ ................................ .................... 45 Comp uterized Tasks ................................ ................................ ........................ 47 Electrophysiology Recordings ................................ ................................ .......... 50 Data Analysis ................................ ................................ ................................ .......... 51 Participant Characteristics ................................ ................................ ................ 51 Behavioral Analyses ................................ ................................ ......................... 51 ERPs ................................ ................................ ................................ ................ 52 Interpersonal Difficulties ................................ ................................ ................... 55 General Considerations ................................ ................................ .................... 55 3 RESULTS ................................ ................................ ................................ ............... 59 Descriptive Variables ................................ ................................ .............................. 59 Neurobehavioral Measures of Task Per formance ................................ ................... 61 Simple Reaction Time Task ................................ ................................ .............. 61 Sex Judgement Task ................................ ................................ ........................ 61 Emotion Ju dgement Task ................................ ................................ ................. 62 Interpersonal Functioning ................................ ................................ ....................... 64
7 4 DISCUSSION ................................ ................................ ................................ ......... 76 Limitations ................................ ................................ ................................ ............... 85 Conclusions ................................ ................................ ................................ ............ 87 LIST OF REFERENCES ................................ ................................ ............................... 89 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 101
8 LIST OF TABLES Table page 2 1 Screening Measures and Exclusionary Cutoffs ................................ .................. 57 3 1 Demographic and Affective Measures ................................ ................................ 66 3 2 Alcohol Related Measures (Alcohol Dependent Participants) ............................ 67 3 3 Sex Judgement and Emotion Judgem ent: Task Performa nce ............................ 68 3 4 Sex Judgement Task: Electrophysiological Results ................................ ........... 69 3 5 Emotion Judgement Task: Electrophysiological Results ................................ .... 70 3 6 Inventory of Interpersonal Problems: Tota l and Subscale Scores ...................... 71
9 LIST OF FIGURES Figure page 2 1 Sex Judgement Task: Illu stration of Sex Morphed Stimuli ................................ .. 57 2 2 Emotion Judgement Task: Illustration of Morphed Emotional Facial Expression Stimuli ................................ ................................ .............................. 58 3 1 Emotion Judgement T ask Accuracy ................................ ................................ ... 72 3 2 Emotion Judgement Task P3 Amplitude ................................ ............................. 73 3 3 Gran d Average EEG Waveforms: Emoti on Judgement Task ( Electrode PZ ) ..... 74 3 4 Grand Average EEG Waveforms : Emoti on Judgement Task ( Electrode T6 ) ..... 75
10 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 A NEUROBEHAVIORAL INVESTIGATION OF EMOTION PROCESSING AND INTERPERSONAL FUNCTIONING IN ALCOHOLISM By Lauren A. Hoffman May 2017 Chair: Sara Jo Nixon Major: Psychology A substantial literature has d etaile d neurobehavioral compromise in alcoholism. Alt hough c lassically defined cognitive processes are predominantly assessed, a growing literature is addressing emotion processing. T he study of emotional facial expression (EF E) recognition has gained particular momentum and implicates EFE processing in treatment outcome s B ehavioral investigations conducted to date suggest an alcohol related EFE identification deficit. R esearch has also begun to elucidate the neural systems/processes underlying compromise Still, the precise nature, particular ly its valence/emotion specificity, require s clarification. Given the importance of facial expression in non verbal communication, t he relationship between EFE processing and interpersonal problems (IP) also merits further i nquiry The current study e xamined neurobehavioral indices of EFE processing and their relationship to IP in alcohol d ependent treatment seekers (AD) and healthy controls (CC). Thirty four ADs and thirty nine CCs completed an emotion judgement task (EJT) requiring discrimination between happy, angry, and sad EFEs. A sex judgement task (SJT; neut ral expression male/female face s ) was administered to strengthen EJT
11 interpretations. Behavioral (accuracy reaction time) and neurophysiological (event related potentials; ERP) responses were assessed for both tasks. ERP measures included amplitude and latency of the N170 and later occurring P3. IP was assessed with the Inventory of Interpersonal Problems 64 ( IIP 64 ) EJT, SJT, and IIP 64 measures were analyzed for gr oup dif feren ces EJT and IIP 64 were investigated for w ithin group correlations EJT analyses revealed somewhat less accurate performance and lower P3 amplitudes in ADs than CCs, across all emotions. Although no gro up by emotion interaction s emerged examination of group means suggested a more pronounced difference for anger. IIP 64 analyses indicated greater interpersonal difficulty among ADs. Significant correlations were specific to ADs and anger; poorer accuracy correlated with greater interpersonal difficulty EJ T reaction time, P3 latency, N170 characteristics and all SJT measures yielded non sign ificant group related outcomes. Findings s ug gest atypical EFE processing in ADs including subtle behavioral deficiencies and altered neural activity at later p rocessing stages that are specific to the evaluation of emotionally laden information and generalizable to positive / negative valence. Moreover alcohol associated emotion processing difficulties and IP appear to be related. Preliminary data also suggest an enhanced role of anger in this relationship and neurobehavioral compromise. Further investigation is needed to confirm and extend these findings
12 CHAPTER 1 IN TRODUC TION P sychoactive substance use and associated disorders have been a significant source of individual and societal burden for centuries Of the many psychoactive substances with the capacity to alter perceptive, affective, and cognitive processes, alcohol is the most widely used (excluding caffeine; Ju lien et al., 2010) and continues to pose a significant public health problem in many countries. In the United States, approximately 57% (136.8 million) of Americans age 18 and older are current (i.e. past 30 days) drinkers (Center for Behavioral Health Sta tistics and Quality (CBHSQ), 2015 a ). Among this population, an estimated 59.4 million and 16.1 million individuals meet criteria for current binge (binge 15 a ). The misuse of alcohol is associated with a host of physical and mental health problems and remains a prominent source of economic burden, morbidity and mortality. Alcohol misuse is responsible for over 200 billion dollars in annual costs (Sacks et a l., 2015) and is currently ranked as the third leading cause of death. Moreover, harmful alcohol use directly accounts for more than 25 chronic diseases and conditions (National Institute on Alcohol Abuse and Alcoholism (NIAAA) 2014), including liver cirr hosis, pancreatitis, and Alcohol Use Disorders (AUDs) the latter of which being entirely alcohol attributable ( World Health Organization (WHO), 2014 ). Longitudinal data collected via the National Epidemiologic Survey on Alcohol and Related Conditions sugg est that approximately 22.7% of alcohol users will become alcohol dependent at some point during their lifetime (Galanter et al., 2014). I n the United States alone an estimated 16.3 million American adults currently suffer from an AUD (Substance Abuse
13 and Mental Health Services Administration (SAMHSA), 2014). Accordingly, alcohol accounts for the majority of substance use disorders (CBHSQ, 2015 b ) and is the leading reason for receipt of substance use treatment (SAMHSA, 2013). Alcohol Use Disorders Defined AUDs are characterized by a well defined set of criteria that have been established and periodically revised by the medical and scientific community. Among the various diagnostic tools available, two conventional sets of criteria are most commonly used to Diagnostic and Statistical Manual for Mental Disorders (DSM) and 2) the World Health in terminology and criter ia exists between these diagnostic tools and within their respective versions (e.g. Alcohol Dependence in DSM IV and ICD 10 vs. moderate / severe Alcohol Use Disorder in DSM 5), broad similarities permit a common framework. Recent vers ions of the DSM (DSM IV (1994) and DSM 5 (2013)) and ICD (ICD 10 (1993)) are homogeneously comprised of AUD criteria that address : 1) tolerance (requiring a greater amount of alcohol to achieve the same effect), 2) withdrawal (e.g., nausea, tremors, or sweating in the abs ence of alcohol), 3) impaired control of drinking behaviors, 4) neglect of important activities, 5) excess time dedicated to alcohol related activities, and 6) continued use despite physiological and psychological consequences (Hasin, 2003 ; NIAAA, 2016; Am erican Psychological Association 2014). Therefore, AUD can be broadly characterized as regular and excessive alcohol consumption despite negative consequences, a loss of control over drinking behaviors, and the likely presence of physiological symptoms A lthough the ICD has both clinical and scientific utility the majority of the scientific literature relevant to the current investigation, and
14 subsequently cited below concern s Alcohol Dependent populations per DSM IV criteria. Given that variance between and within these diagnostic tools is not a primary focus of the current report, diagnoses determined via all of these tools are generally referred to throughout the Intr oduction and Discussion for ease of communication Given the nature of the available literature, these terms will principally denote alcohol dependence Unless otherwise stated, all alcoholic participants in the studies noted hereafter have completed detox ification and are typically in or beyond their third week of sobriety. Neurobehavioral Correlates of Alcohol Use Disorders Prevention and successful treatment of AUDs partially relies on the thorough characterization of predictors, correlates, and conseque nces of harmful alcohol use behaviors. Given the prevalence of and individual/societal burden caused by AUDs, it is no surprise that substantial resources have been dedicated to better understand them. The deleterious effects of chronic heavy alcohol use h ave been widely investigated over the last several decades, offering significant insight to the neurobehavioral correlates of alcoholism. S tudies have revealed a host of alcohol related structural (Fein et al., 2002; Makris et al., 2008), functional (Herma nn et al., 2007; Jung et al., 2014; Pfefferbaum et al., 2001 ), and cognitive abnormalities (Sullivan et al., 2000; Chanraud et al., 2007). Detoxified alcoholics frequently show reductions in white (Agartz et al., 2003 a ; Demirakca et al., 2011; Pfefferbaum et al., 2006) and grey matter (Agartz et al., 2003 a ; Demirakca et al., 2011; Chanraud et al., 2009), enlarged ventricles, and widened sulci (Mann et al., 2005; Wobrock et al., 2009 ; see Oscar Berman et al., 2014 for a recent review ). This neurostructural a trophy has been observed in both whole brain and region specific manners, with the frontal lobes, limbic structures (e.g., amygdala,
15 cingulate cortex, hippocampus, and temporal lobe gyri), cerebellum, and their respective connections showing particular vul ( Bates et al., 2013 ). Furthermore, studies implementing functional magnetic resonance imaging (fMRI), positron emission tomography (PET), and electroencephalography (EEG) provide evidence of atypical cerebral blood flow (e.g., Bagga et al., 2014; Mller Oehring et al., 2013), brain metabolism (e.g., Dao Castellana et al., 1998), and neurophysiology ( e.g., Namkoong et al., 2004; Roopesh et al., 2010), respectively (Oscar Berman et al., 2014) At the behavioral level, domain specific compromise after chronic heavy alcohol consumption has been extensively detailed for a variety of classically defined cognitive abilities. Although not all alcoholics exhibit cognitive insult during abstinence (i.e a period in which alcohol is not consumed), mild to severe impairment can be observed in approximately 30% 80% of this population after detoxification (Bates et al., 2002 a ). Alcohol related deficits span a broad range of cognitive functions. Whereas auto matic and overlearned processes (e.g., crystalized intelligence) are typically spared, notable impairments are observed for more complex and effortful processes (Nixon & Glenn, 1995; Stavro et al., 2013). More specifically, an extensive literature has docu mented deficient visuospatial, planning, and problem solving skills within the domain of fluid ability, as well as impaired prospective memory, episodic memory, and novel learning skills within the domain of learning / memory (Bates et al., 2013; Pitel et al., 2014) Moreover, executive functions such as response inhibition, mental flexibility, self monitoring, and working memory are suggested to be particularly vulnerable to alcohol related compromise. Other cognitive functions commonly affected in populat ions with
16 AUD include psychomotor skills, information processing speed, and attention ( Bates et al., 2013 ; Pitel et al., 2014 ). Many of these impairments persist throughout prolonged abstinence (Fein et al., 1990; Stavro et al., 2013); however, partial rec overy of cognitive function (Fein et al., 2006) and brain atrophy (Agartz et al., 2003 b ; Demirakca et al., 2011; Gazdzinski et al., 2010) is reported (Oscar Berman et al., 2014) In an effort to explain alcohol related cognitive insult, several theories of effects on neural structure and function have been proposed. Among these theories, the frontal lobe hypothesis suggest s that anterior neural structures are most susceptible to tion of studies showing diminished neurobehavioral functioning among alcoholics in domains thought to be subserved by frontal structures (e.g., executive functions ; Parsons, 1987; Parsons & Leber, 1982 ). However, the frontal lobe hypothesis does not fully account for the heterogeneity or lack of frontal lobe dysfunction seen in some alcohol related investigations and fails to explain commonly observed deficits in other neural regions and their associated functions. A lternative ly, the lateralization hypothes is proposes that right hemisphere structures/functions are disproportionately vulnerable. Support for this hypothesis stems from reports of deficient right hemisphere lateralized non verbal functions in alcoholics, with left hemisphere lateralized verbal a bilities left relatively intact (Parsons, 1987; Parsons & Leber, 1982) Despite support for the lateralization hypothesis this pattern of impairment is not always observed ( Stavro et al., 2013 ). A third theory suggests that alcoholism affects the brain in a global manner, inducing mild disruptions to multiple brain areas that variably contribute to functional neural systems (Parsons & Leber, 1982)
17 garnered considerable support over the years (e.g ., Beatty et al., 2000; Stavro et al., 2013; Tivis et al., 1995). Consequently, an attempt to identify focal lesions in alcoholics has been largely abandoned. This shift in research has led to the widespread investigation of alcohol related processing defi cits, which has in turn, suggested that cognitive impairment may be the result of inefficient processing (Nixon & Bowlby, 1996) and limited cognitive resources (Fama et al., 2004; Smith & Oscar Berman, 1992). Characterization of AUDs and the neurobeh avioral systems they affect has considerable potential to inform effective treatment and patient recovery. Despite a rich history of research concerning the neurobehavioral correlates of alcoholism systematic investigation has been somewhat limited to cla ssically defined cognitive processes. Among the functional domains that have received relatively less attention and merit add itional investigation, emotion processing has significant potential to contribute to a more comprehensive understanding of AUDs. Em otion Processing and Alcohol Use Disorders The perception, interpretation, and regulation of emotional states are essential survival mechanisms that influence cognitive processes (e.g., decision making) and subsequent behaviors (Adolphs, 2002; Damasio, 199 9). Among their many functional roles, these affect related components or emotional processes, serve as key facilitators of social communication. More specifically, effective processing of i nternal and external (auditory and visual) emotional cues promotes socially acceptable and situation appropriate interpersonal interactions (Adolphs, 2002; Damasio, 1999). The study of emotion processing in populations with AUD suggests dysfunction in a variety of affect related areas. For example, alcoholics commonly exhibit alexithymia, a condition characterized by the reduced ability to experience, identify, and
18 describe personal emotional states (Kopera et al., 2014). Blunted affect, disinhibition, altered reward processing, antisocial personality traits, and difficulty interpreting the verbal and non verbal affective cues of others have also been noted ( Oscar Berman & 2007 ; Parsons et al., 1987 ). Altered emotional functioning is thought to play a primary role in AUD developm ent and disease progression ( Clark et al., 2007 ; Zywiak et al., 2003 ). Relatedly, e motion processing deficits can negatively impact psychosocial and interpersonal domains of daily living, for which problems are commonly reported in alcoholics and serve as important determinants of recovery (Marlatt, 1996) One component of emotion processing that plays a key role in interpersonal functioning and has received increased attention in AUD focused research over the last decade is emotional facial expression (EFE ) identification ( Uekermann & Daum, 2008 ). A growing body of litera ture suggests that alcoholism is associated with a decreased ability to accurately identify and interpret the EFEs of others. Facial expression is a particularly important component of non verbal communication of affective states ( Eimer & Holmes, 2002 ). Although some controversy exists as to which facial expressions are categorically defined as emotions the scientific community has largely agreed that f aces code for six basic emotions: happ iness, anger, sadness, fear, surprise, and disgust ( Uekermann & Daum, 2008 ). Altered processing of emotional facial cues can lead to misinterpretation of outcomes. For e xample, misperception of anger might result in altercation and interpersonal conflict. Furthermore, impaired emotion processing and resultant interpersonal difficulties may promote avoidance of social situations. Importantly, social
19 isolation has the poten tial to hinder treatment efforts and increase risk of relapse in recovering alcoholics ( Uekermann & Daum, 2008 ). In fact, p oorer EFE identification has exhibited a direct relationship with increased likelihood of treatment drop out ( Foisy et al., 2007 a ) and a greater number of medical detoxifications ( Townshend & Duka, 2003 ). Clearly the investigation of emotion processing in alcoholism has important and clinically relevant implications. C haracterizing the nature and extent of alcohol related EFE process ing deficits and the way they relate to interpersonal difficulties, might better inform treatment efforts and ultimately facilitate AUD recovery. The study of EFE processing among abstinent alcoholics gained momentum in arch has been conducted in Belgium by a single group of investigators and has largely concerned behaviorally circumscribed markers of EFE processing. These studies, along with the few originating from other countries (e.g., England, Germany, Italy, United States), most commonly assess ed behavioral outcomes by measuring accuracy of EFE identification and/or perception of emotional intensity depicted in facial expressions of emotion (i.e. intensity ratings ; Bora & Zorlu, 2017 ). Reaction time has been investig ated to a lesser extent, as many behavioral studies have either elected not to restrict stimulus presentation and response windows or to omit reaction time findings when response windows have upper limits. Some of these investigations have used EFE identif ication tasks that consist of face stimuli express ing full scale emotion (i.e. maximum (100%) intensity). Others utilize faces that are more likely to be encountered in real world situations and, therefore, have greater ecological validity (Kornreich et al ., 2003) In t hese cases, faces either depict 1) a lesser degree/intensity of emotion (i.e. EFE morphed with neutral facial expression to depict a
20 proportion of the intended emotion; e.g., 30% happy / 70% neutral) or 2) mixed emotions (i.e. two different E FEs morphed with one another, resulting in one face that expresses a proportion of each emotion; e.g., 30% happy / 70% fear). Previous behavioral investigatio ns of alcoholics, relative to control samples, have revealed less accurate EFE identification for maximum intensity expressions (Quaglino et al., 2015; Carmona Perera et al., 2014 ; Kornreich et al., 2013 ; Maurage et al., 2008 a ; Frigerio et al ., 2002 ), emotion/neutral morphed expressions (Philippot et al., 1999 ; Kornreich et al., 2003 ; Kornreich et al., 2001 a ; Foisy et al., 2007 a ; Kornreich et al., 2001 b ), and mixed emotion expressions (Acharya & Dolan 2012; Townshend & Duka 2003 ; Maurage et al., 2007 a ) Findings also suggest that alcoholics have a tendency to overestimate emotional intensity in facial expressions (Townshend & Duka 2003; Philippot et al., 1999 ; Kornreich et al., 2001 a ; Maurage et al., 20 09 ; Kornreich et al., 2013 ; Kornreich et al., 2001 b ; Foisy et al., 2007 a ) and might require greater intensity of emotional expression for accurate EFE identificat ion (Frigerio et al., 2002 ). Although response latency is less frequently explored alcohol related reaction time delays in EFE identification h ave been previously noted in a handful of reports ( e.g., Maurage et al., 2007 a ). While g eneral psychomotor slowing is occasionally observed in abstinent alcoholics regardless of neurobehavioral demands, some studies have shown emotion specific reaction tim e delays without generalizability to the identification of other facial attributes (e.g., sex, race, age; Maurage et al., 2008 a ; Foisy et al., 2007 b ). Importantly, other investigations provide evidence that alcohol related impairments in EFE identificatio n are not due to a reduced ability to label emotions (i.e. semantic deficit) but
21 rather, appear to be attributable to deficient perceptual and/or top down integrative processes (Kornreich et al., 2016). Despite substantial evidence for an emotion processing impairment behavioral deficits in EFE identification are not always observed among alcoholics (see Uekermann et al., 2005 for null findings) Moreover, when impaired performance is noted the precise nature of this deficit has yielded equivoca l findings. The greatest inconsistencies are observed for valence specific outcomes. Some investigations suggest that heavy chronic alcohol consumption might induce EFE processing deficits specific to negative ly valence d emotions (i.e., unpleasant; e.g., s ad), with positively valenced emotion (i.e. pleasant; e.g., happy) identification spared. More specifically, t hese studies reveal less accurate identification of negative valence among recently abstinent alcoholics including but not limited to, anger and /or sadness, and a preserved ability to identify happiness (Quaglino et al., 2015; Acharya & Dolan 2012 ; Car mona Perera et al., 2014; Frigerio et al., 2002 ; Townshend & Duka 2003 ; Kornreich et al., 201 3 ) and/or surprise (Acharya & Dolan 2012 ; Car mona Pe rera et al., 2014 ; Townshend & Duka 2003 ). Investigations that demonstrate alcohol related overestimation of the emotional intensity depicted in EFEs have also suggested negative valence specificity In other words, alcoholics might overestimate the amoun t of negative, rather than positive, valance exp ressed in all EFEs (positively and negatively valenced expressions; Foisy et al., 2007 a ), with some studies showing the particular overestimation of an ger (Frigerio et al., 2002 ; Townshend & Duka 2003; Philippot et al., 1999; Maurage et al., 2009 ).
22 Despite evidence for a valence dependent deficit a number of other investigations have observed impaired EFE identification for both positively and negatively valenced stimuli (Maurage et al., 2008 a ; Philippo t et al., 1999; Kornreich et al., 2001 a ; Kornreich et al., 2003; Maurage et al., 2007 a ; Kornreich et al., 2001 b ). Importantly, when impaired EFE identification of positive valence has been reported, it has been consistently accompanied by an additional def icit for at least one negatively valenced emotion. Therefore, EFE processing impairments might be more pronounced for negative ly valence d emotions Although scientific evidence generally supports a behavioral deficit for negatively valenced EFEs, inconsist encies have prevented conclusions specificity to emotions within this valence category For example, investigation of recently abstinent alcoholics, as compared to community controls, has revealed less accurate identification of ang er, but preserved recognition of fear (Kornreich et al., 2001 a ). Some studies have shown support for this finding and reveal additional impairments for the identification of sad and disgusted EFEs (Acharya & Dolan 2012; Philippot et al., 1999; Kornreich e t al., 2001 b ). However, several reports note contradictory results. Alcoholics have alternatively exhibited a preserved ability to identify anger and sadness but less accurate identification of fear This alcohol related deficit for fear has been observed in both the presence (Carmona Perera et al., 2014) and absence ( Townshend and Duka, 2003) of impaired disgust recognition. O ther reports have noted unimpaired fear recognition along with typical anger and sadness identification, and otherwise suggest a de ficit specific to disgust (Quaglino et al., 2015) T he preserved identification of anger and sadness observed in these studies has been
23 reported by another team of researchers (Kornreich et al., 2013) Yet, in contrast to these findings and those noted abo ve, alcoholics have also exhibited less accurate identification of sadness with unimpaired recognition of anger (Frigerio et al., 2002 ). In an attempt to clarify inconsistent findings, a recent m eta analysis (Bora & Zorlu, 2017 ) estimated mean effect sizes across studies of facial emotion recognition impairments in abstinent alcoholics. The largest effect sizes were observed for anger and disgust, suggesting that the alcohol related EFE identification deficit might be most pronounced for these emotions. How ever, this meta analysis only included 12 studies of EFE processing (analyses were limited to behavioral studies of four or more maximum intensity or emotion/neutral morphed expressions that reported sufficient information for calculation of effect sizes a nd standard errors), reinforcing the need for further investigation. Taken together, these highly variable outcomes currently preclude conclusions regarding an emotion specific deficit for negatively valenced EFEs. Contributing to this confusion, some inve stigations have revealed global rather than valence or emotion specific, EFE identification impairments. In these studies, abstinent alcoholics show poor identification across all emotions of interest. Global deficits have been noted with simple two cho ice paradigms using happy angry and sad EFEs (Maurage et al., 200 8 b ), as well as with more complex response demands and additional EFE stimuli (e.g., evaluation of 4 EFE stimuli and respon se options for 8 emotion expression s ; Kornreich et al., 2003 ). Clearly, behavioral deficits in the context of emotion /valence specific findings require further investigation to elucidate the precise nature of AUD related impairment. Assessment of EFE processing at the neural level might provide additional insight and help to clarify inconsistent findings.
24 Behavioral tests provide gross measures of function, which offer insight to the external manifestations of complex neural processes. Subtle deficits and the use of alternate strategies might not be detected by th ese assessments when administered in a laboratory setting. Furthermore, the neural structures and specific processes implicated in typical and atypical EFE identification cannot be fully delineated with behavioral measures. Therefore, examining the neural correlates of EFE processing impairments may provide a more complete understanding of dysfunction in AUDs. To date, the characterization of alcohol related EFE processing deficits at the neural level remains relatively understudied ( ). Among existing reports fMRI investigations have suggest ed altered activation using a variety of EFE based tasks Although behavioral impairment is not always observed with these assessments (potentially due to relatively low task complexity ; Charlet et al ., 2014; Salloum et al., 2007 ), atypical neural activation is consistently reported. Relative to non alcoholic controls, abstinent alcoholics have exhibited reduced neural activation while performing 12), intensity evaluation (Salloum et al., 2007), valence (positive vs. negative) cate gorization (Maurage et al., 2013 ), and implicit emotion processing (Marinkovic et al., 2009; Charlet et al., 2014 ). Brain regions showing deficient activ ation in response to the EFE evaluation include the anterior cingulate cortex (Salloum et al., 2007), orbitofrontal cortex, insula ntal gyrus (Maurage et al., 2013 ), amygdala, and hippocampus (Marinkovic et al., 2009). Com pensatory neural activity has also been suggested, in the absence of significant neural or behavioral deficit s for regions
25 associated with higher order functions (e.g., prefrontal cortex; Charlet et al., 2014 ). Despite between study variability in methodo logy and region specific findings, many of these structures have been broadly implicated in EFE processing among healthy non alcoholic populations (Adolphs, 2002). For example, studies that utilize imaging techniques, single cell recordings, and focal brai n lesioned populations have shown that the evaluation of facial features and expressions involve s the orbitofrontal cortex, amygdala, anterior cingulate cortex, parietal cortex, inferior parietal cortex, occipito and infero temporal cortex, as well as ot her frontal structures ( Adolphs et al., 1996; Blair et al., 1999; Fried et al., 1997 ). Clearly, EFE identification recruits a host of brain regions. These regions contribute to several component processes of EFE analysis and do so i n a temporally dynamic f ashion. Although impractical to assign a discrete neural structure to a single component process (Adolphs, 2002) neural network theories of EFE processing suggest temporally dependent structural contributions. E vidence based theories propose that EFE proc essing begins with structural encoding of facial features for the formation of crude stimulus representations and categorization s (e.g ., face vs. other object). This rapid and early perceptual processing typically occurs within approxima tely 100 ms from st imulus onset. It is suggested to involve brain regions comprising the core visual system, including e arly visual cortices (e.g., V1 and V2) and subcortical structures such as the superior colliculus and pulvinar thalamus M ore anterior regions (including v isual association cortices) are thought to be subsequently recruited for further perceptual processing By approximately 170 ms, information processing via posterior temporal cortices and the fusiform gyrus purportedly allow for the formation of
26 detailed structural representations. Representations of animated (e.g., expression) and static (e.g., identity) facial features, respectively contained in the superior temporal sulcus and lateral fusiform gyrus, are collectively provided to higher order co gnitive systems. During the later stages of EFE processing (~300 ms post stimulus onset) the anterior cingulate cortex, prefrontal cortex, and somatosensory regions are allegedly recruited by this dynamic network for the formation of conscious EFE stimulu s representations, control of thoughts and behaviors, and elicitation of accompanying internal states (e.g., associated feelings). Parallel neural activation of the amygdala, orbitofrontal cortex, and ventral striatum is likely to modulate early and late E FE processing stages. Neuronal projections to and from these structures can influence attentional allocation for facial features, refine stimulus representations and classification according to their emotional significance, modulate retrieval of stored emo tion related knowledge, and induce an emotional response ( for a review of these concepts, see: Ado lphs, 2002; Eimer & Holmes, 2007 ). In sum, neural network theories propose that EFE processing involves the fluid and collective effort of several brain regio ns that have variable and shared contributions to temporally defined component processes. One way to measure the collective contributions of this complex neural network is through the use of techniques with superior temporal resolution. Electrophysiologica l techniques are particularly well suited for the investigation of discrete, temporally defined neural processes that arise from the summated activity of neural systems and constitute global behaviors. Electroencephalography (EEG) allows for the recording of synchronous post synaptic electrical brain activity at the level of the scalp. Event related potentials (ERPs)
27 can be extracted and averaged from continuous EEG recordings to reveal voltage oscillations, time locked to specific internal and ex ternal eve nts (Luck et al., 2014 ). ERP waveforms are thought to reflect certain psychological processes and are typically measured with amplitude and latency (Porjesz & Begleiter, 1995; See Coles et al., 1991 for a more thorough review on ERPs). ERP components are t ypica lly named according to the direction of their d eflection (i.e. P for positive and N for negative) and the approximate time at which they peak from stimulus onset (e.g., P100 is a positive going waveform peaking approximatel y 100 ms after stimulus ons et). Researchers have largely relied on neurophysiological measures to better characterize the component processes of a given function. Whereas behavioral testing provides gross measures of cognitive function, neurophysiology allows for the observation of neural activity on a millisecond time course. Thus, individual components, associated with specific psychological/cognitive processes, can be separated from global behaviors and assessed as distinct stages (Tipper, 2004). ERP components can be broadly cate gorized into two general stages: 1) early exogenous processes that are stimulus driven and largel y automatic (early latency ; e.g., 50 100 ms ) and 2) later endogenous processes that are associated with higher order functions, such as conscious appraisal, th e formation of complete stimulus representations, and decisional processes (mid and late latency; e.g., 300 500 ms; Fein et al., 2010 ; Luck, 2014 ). Evidence suggests that both earlier and later ERP components, reflecting various processing stages, can be influenced by emotional information portrayed in a stimulus. Mid latency components have exhibited sensitivity for detecting emotional versus neutral facial expressions, whereas later latency components have show n
28 decipherable activation between EFEs (e.g ., happy vs. fearful; For a review, see Hajcak et al., 2012). ERP aberrations suggest atypical activation of neural networks that contribute to a given process. Similar to neuroimaging, atypical ERP measures may be observed in the absence of behavioral imp airment thus, providing a more sensitive index of alcohol related functional abnormalities. To date, relatively few ERP investigations of EFE processing have been conducted in the AUD population. Among them, significant varia bility exists in the tasks used. Given that ERP elicitation and interpretation are highly task dependent, components of interest have also differed. Nonetheless, ERP investigations have begun to shed light on the neural correlates that underlie EFE related behavioral impairments in recently abstinent alcoholics. The current ly available literature includes ERP studies of both implicit and explicit emotion processing. I mplicit EFE processing has been previously examined with emotio nal oddball paradigms. Although t ask de mands do not call for explicit evaluation of emotion al expression, they require the processing of stimuli that contain emotional information 2012) Oddball tasks are simple and widely implemented paradigms used in the broader ERP literature to investigate attention mediated processing. In an oddball task, rare target stimuli are presented within a series of frequent (i.e. standard) non targets (Luck et al., 2014) For the emotion variation of this task, simple stimuli (e.g., objects and shap es) typically used in a classical paradigm are substituted with stimuli that contain emotional content, such as EFEs. T wo investigations have used the emotional oddball task to assess earlier and later stages of EFE processing among recently abstinent alco holics. These studies reveal behavioral impairment, as well as
29 electrophysiological markers of deficient implicit emotion processing among individuals with AUD, relative to healthy community controls However, the specific processing stages implicated in b ehavioral impairment differ between studies Concerning earlier processing stages, b oth investigations assessed the N170 in response to facial expression stimuli. The N170 is a n ERP component that shows preferential activation for human faces and is thought to reflect higher level perceptual processing stages such as stimulus encoding, which allows for the formation of complete stimulus representations (Bentin et al., 1996 ; Rossion & Jacques, 2011 ). O ne study demonstrated emotion and valence specif ic alterations of the N170 among recently abstinent alcoholics compared to healthy control participants. More specifically, alcohol related N170 latency delays were more pronounced for sad and fearful EFEs than neutral and happy expressions presented at v arying emotional intensities (Maurage et al., 2007 b ). In contrast, the same group of researchers observed typical N170 component characteristics among a different group of alcoholics when mixed emotion expressions of anger and disgust were presented under varying oddball task conditions ( Maurag e et al., 2008 c ) Both of these studies also examined the later occurring P300 (i.e. P3) component, which is generally associated with decisional processing prior to initiation of a motor response (e.g., button press ; Hansenne, 2006; Polich, 2004) and is thought to reflect the amount of information extracted from a given stimulus (Donchin et al., 1986; Campanella et al., 2010; Campanella et al., 2009). Altered P3 activation was unanimously reported across investigatio ns. However, Maurage and colleagues observed delayed latency and attenuated amplitude of the P3 across all facial
30 expressions (neutral and emotional) in their earlier investigation ( Maurage et al., 2007 b ) whereas their later report suggested an anger specific processing deficit among alcoholics ( Maurage et al., 2008 c ) Regarding the latter finding, h ealthy controls showed neural activation suggestive of enhanced processing of, and greater difficulty disengaging neural resources from, anger, and a lcoholics failed to show this beneficial adaptive response (assuming anger is a biologically relevant cue for danger and altercation) Atypical component characteristics were accompanied by delayed reaction times in both investigations with alcoholics sho wing delays that were either specific to sad EFEs (Maurage et al., 2007 b ) or generalized to all EFEs ( Maurage et al., 2008 c ) Taken together it is not entirely clear if the implicit emotion processing deficit presents for alcoholism at higher level perc eptual processing stages (i.e. N170). However, investigation of P3 component activation seems to consistently reveal atypical characteristics among alcoholics suggesting altered neural activation, at least at decisional processing stages, that manifests as behavioral impairment. Still, conflicting findings and varying task demands preclude strong conclusions regarding emotion and valence specificity. Although the oddball task is a popular and useful paradigm for electrophysiological research, relatively low task complexity often yields ceiling effects for behavioral measures of accuracy Furthermore, task demands do not necessarily elicit explicit emotion processing, for which identification and categorization of emotions is necessary. A handful of investiga tions have assessed explicit emotion processing, in which tasks demand e d discrimination between two or three EFEs. In other words,
31 participants were required to identify which of two (or three) emotions was expressed by a given face stimulus. Given that a markedly limited number of ERP studies have assessed explicit EFE processing in recently abstinent alcoholics methodology, EFE stimuli, components of interest, and study outcomes are highly variable between them. One of these investigations (Maurage et al ., 2008 b ) examined emotion processing in a lcoholics using an emotion judgement task that required the discrimination of happy, angry, and sad EFEs, morphed with neutral expressions to depict varying degrees of each expression (35%, 65%, 95%). Relative to h ealthy community controls, abstinent alcoholics exhibited delayed reaction times and atypical ERP component characteristics across all emotions P100, N170, and P3 latency delays as well as attenuated P3 amplitude were noted The administration of control tasks, including a sex judgement task and a simple reaction time task, further reveal ed that atypical P3 elicitation and reaction time delays were specific to emotion judgement Taken together, these findings might suggest an EFE processing deficit that i s valence general, attributable to deficient processing at later decisional stages and specific to the evaluation of emotionally laden information Although this study holds significant merit interpretation of alcohol related findings in this report are somewhat obscured by the additional assessment of comorbid depression, which was the primary focus of investigation. A different pattern of results was observed by the same investigative team in an assessment of unimodal versus multimodal (visual and au di tory emotional cues) emotion processing. In unimodal visual conditions, Maurage and colleagues (2008 d ) used maximum intensity happy and angry faces to investigate EFE identification.
32 Comparison of recently abstinent alcoholics and healthy controls revealed alcohol related performance deficits a cross all emotional expressions. Although their emphasis on multimodal investigation makes it somewhat difficult to delineate findings explicitly pertaining to emotion processing in the visual modality, the report clearly communicates P3 amplitude attenuation among alcoholics that is specific to angry EFEs and the absence of earlier ne ural processing (P100 and N170) abnormalities. Consistent with their previous investig ation, findings suggest a reduced ability to correctly identify emotion al facial expressions among alcoholics and atypical neural activation at decisional processing stages. In contrast, these results point to altered neural activity that might be valence dependent However, given that the study was designed to address multimodal processing and stimulus presentation was restricted to two emotions, no formal conclusions can be made as of yet. Nonetheless, the results obtained from these preliminary experimen ts provide a useful foundation for future investigations of AUD. E xplicit emotion processing has also been studied in long months) alcoholics suggesting persistent b ehavioral impairment and atypical neural activation Fein and colleagu es (2010 ) demonstrated these neurobehavioral deficits with maximum intensity happy and sad facial expression s, for which l ong term abstinent alcoholics showed longer reaction times and delayed latency of the P160, a component elicited by attended and unatt ended facial stimuli. In this particular study neurobehavioral deficits did not differ by emotion type and measures of accuracy, P160 amplitude, and P3 characteristics did not show any alcohol related compromise These findings might suggest that EFE proc essing deficits persist in long term abstinence yet,
33 neural inefficiency might present differently than that which is observed in recent abstinence. Importantly, e vidence exists for some improvement in EFE processing with long term abstinence (Kornreich e t al., 2001 b ) Abstinence dependent recovery of emotion processing abilities might suggest that impairment is at least partially attributable to the deleterious effects of chronic heavy alcohol use and not entirely a function of an AUD related predisposit ion ( Uekermann & Daum, 2008 ). Therefore, detailed characterization of emotion processing in alcoholism has the potential to inform treatment efforts and facilitate recovery. The inconsistencies noted above emphasize the need for further inquiry as to whether or not differential deficits exist according to valence and emotion type and to better characterize the component processes implicated The relationship between EFE processing and interpersonal problems also merits greater attention. Kornreich and colleagues (2002) have previously reported a relationship between EFE processing and interpersonal functioning, as well as the potential for interpersonal difficulties to increase risk of relapse during early and mid term abstinence Despite these compel ling preliminary data direct investigation of the association between interpe rsonal functioning and emotion processing is surprisingly scarce in the AUD relevant literature. Among AUD focused investigations of EFE processing interpersonal problems have been predominantly assessed by Maurage and colleagues. In their investigations, i nterpersonal functioning was evaluated with the Inventory of Interpersonal Proble ms (IIP; Horowitz et al., 1988). The IIP assesses the quantity a nd quality of social interactions and personal relationships providing measures of overall interpersonal
34 problems (overall scores) and difficulty within discrete i nterpersonal domains (sub scale scores). Using this inventory, the investigation of interpers onal functioning in Maurage behavioral and ERP studies (noted above) have yielded inconsistent findings. Some data suggest equivalent overall IIP scores between recently abstinent alcoholics and community controls (Maurage et al., 2007 b 20 07 a 2008 a ). Other assessments have revealed higher IIP scores (i.e. higher levels of interpersonal difficulty) among alcoholics, both overall (Maurage et al., 2008 c 2009 ) and on specific subscales (Maurage et al., 2008 d 2009 ). In one of these reports, Ma urage and colleagues (2009 ) investigated the relationship between interpersonal functioning and emotion identification For this study, a variety of emotionally laden stimuli (i.e., faces, postures, voices, short stories) were presented and participants rated each stimulus on multiple Likert scales to indicate the emotions they believed to be depicted. Average ratings across all emotional stimuli were assessed for correlations with IIP scores, revealing a relationship between emotion processing and the se lf control problems subscale of the IIP among alcoholics ( inventory and subscales detailed in Chapter 2, Methods) More specifically, greater difficulty with self control, in the context of interpersonal relationships, was associated with a n overall tenden cy to judge emotionally laden stimuli as depicting a greater degree of anger and threat. Higher scores on this subscale have also been shown to correlate with greater difficulty identifying endogenous emotional state s ( i.e. self experienced emotions; Maura ge et al., 2008 d ). Importantly, the correlations noted here were specifically observed among alcoholics, and significant group differences (AUD vs. CC) for ERP and behavioral measures have still been observed when between group
35 variability in interpersonal problems are controlled for (Maurage et al., 2008 c ). However, additional research is needed among other investigative teams and AUD populations before any clear conclusions can be made about the association between EFE processing and interpersonal functio ning. Given the suggested relationship between interpersonal difficulties and early recovery outcomes, further characterization can provide significant insight not only to our understanding of emotion processing in AUD populations, but also to our understanding of the components that impact recovery. Summary The prevalence of alcoholism continues to surpass all other substance use disorders ( CBHSQ, 2015 b ) with an estimated 16.3 million American adults meeting criteria for a current alcohol use disorder ( SAMHSA 2014 individual and societal impact, several decades of scientific research have been directed toward better understanding the deleterious effects of chronic heavy alcohol use A substan tial literature has detailed cognitive processing impairments in populations with alcohol use disorders and, more recently, investigations have begun to address the domain of emotion processing. Taken together, preliminary evidence exists for an emotion pr ocessing deficit in alcoholism. Although behavioral impairment is not always observed, s everal studies demonstrate atypical EFE processing including reduced ability to accurately identify emotions and overestimation of the emotional intensity perceived in facial expressions. A growing literature has also begun to elucidate the corresponding neural systems and component processes implicated in alcohol related emotion identification difficulties. Yet the exact nature of this deficit and the factors associat ed with it merit further investigation. In particular, valence and emotion specific findings require clarification. If a deficit exists, and is specific to
36 negatively valenced emotions in general or one emotion in particular (e.g., anger), this can have im portant implications for therapeutic strategies Assessing the relationship between these emotion identification difficulties and various facets of interpersonal functioning might also advanc e our understanding of emotion processing in alcoholism thereby informing treatment efforts aimed at successful rehabilitation.
37 CHAPTER 2 STUDY AIMS AND METHODS Study Aims In an effort to contribute to this growing litera ture and elucidate some of the unresolved issues, the current study examined neurobehavioral ind ices of EFE processing in alcohol dependent treatment seekers, relative to healthy, non alcoholic community controls. Behavioral i nvestigation of EFE processing was conducted using measures of accuracy and reaction time for an emotion judgement task Because EFE processing impairments could be attributable to global visuospatial processing deficits and/or general psychomotor slowing, performance measures were also assessed for cont rol tasks (sex judgement task and simple reaction time task ; detailed in subsequent sections ). Importantly, evidence suggests that performance based deficits observed in abstinent alcoholics may be the result of inefficient processing (Glenn & Parsons, 1990, 1992) Efficiency of the processes that underlie emotion identificati on can be investigated in several ways; our laboratory has used speed accuracy tradeoffs as one way to assess this (e.g., Hoffman et al., 2015; Nixon et al., 2007) With this measure, high levels of accuracy accompanied by fast reaction times reflect effic ient processing (Kaplan, 1988) Trade offs are sometimes observed in abstinent alcoholics such that individuals may slow responses to achieve greater accuracy, or respond more quickly, but at the cost of accurate performance (Glenn & Parsons, 1990, 1992) Thus, in an effort to apply a process oriented approach to these data, behavioral measures were also assessed for speed accuracy trade offs via efficiency ratios (% accuracy / mean reaction time for accurate trials).
38 Consistent with this approach, event re lated potentials (ERPs) were used to investigate the integrity of underlying processes at the neural level. Based on previous research, two endogenously driven components were investigated: 1) N170 to assess neural responses associated with higher order pe rceptual processing of faces, and 2) the P3 to examine processes linked to the decisional stages of stimulus evaluation (here, emotion judgement) The N170 is a negative going waveform that occurs between approximately 130 and 200 ms after stimulus onset a nd is maximal at occipito temporal sites. Multiple cortical regions are thought to contribute to N170 activation, but the lateral portion of the posterior fusiform gyrus and the anterior/middle fusiform gyrus have been suggested as dominant sources (Rossio n & Jacques, 2011). P3 occurs at relatively later processing stages, approximately 300 5 00 ms post stimulus presentation (Campanella et al., 2009 Luck et al., 2014 ). This positive going waveform is maximal at parietal electrodes (Luck, 2005; Knight, 1996) ; the medial temporal lobe and the temporal parietal junction are suggested sources of P3 elicitation (Polich, 20 07). In addition to its association with decisional processing sta ges, P3 component activation might also reflect efficiency of stimulus proces sing (Campanella et al., 2009). Some investigators suggest that P3 component activation is accompanied by widespread neural inhibition of ongoing activity; thereby, increasing the resource s available for further processing of task relevant stimuli (Campane lla et al., 2009). Previous investigations have shown N170 and P3 component sensitivity to emotion (relative to neutral expression; for a review, see Hajcak et al., 2012) and some suggest that activation might be modulated by emotion type. For example, rel ative to happy expressions, sad and angry expressions have elicited delayed N170 (Batty & Taylor,
39 2003) and P3 latency (Maurage et al., 2008 d ), respectively. Recorded during the emotion judgement task and sex judgement task, these components and their corr esponding behavioral measures allowed us to extend a growing literature concerning emotion processing in alcoholism with the following project aims: Aim 1 Compare neurophysiology and behavior in alcohol dependent treatment seekers (ADs) and non alcoholic c ommunity controls (CCs) using an emotion judgement task that required identification of happy, angry, and sad facial expressions. Consistent with previous investigations, we p redicted Hypothesis 1) Emotion processing would be altered in ADs. More specifica lly, we predicted that ADs would exhibit poorer task performance (on at least one behavioral measure) and neurophysiological alterations, including attenuated amplitudes and possibly delayed latencies. Concerning neurophysiological alterations, we asked Em pirical Question 1) whether atypical activation would be observed at earlier perceptual processing stages, later decisional processing stages, or both Furthermore, e vidence from a recent investigation (Maurage et al., 2008 b ) suggests that alcohol related neurophysiological alterations and behavioral deficits in facial processing might be specific to emotion and not generalizable to other domains of face discrimination (e.g., sex, race, age). Therefore, we predicted Hypothesi s 2) Behavioral and/or neurophysiological alterations in ADs would be specific to emotion judgement processing, such that dependent measures would be similar between ADs and CCs when task performance required non emotional face discrimina tion (i.e. sex jud gement task).
40 A primary objective of the current investigation was to shed light on whether or not abstinent alcoholics exhibit a differential deficit across various emotions. Mixed findings regarding the impact of valence and emotion type restricted a suf ficient hypothesis. Therefore, we asked Empirical Question 2) If neurobehavioral indices of atypical emotion processing would be moderated by emotion type (happy vs. angry vs. sad). Aim 2 Examine th e relationship between emotion processing and interpersona l functioning among ADs and CCs. Given preliminary evidence suggesting a relationship between emotional facial expression processing and interpersonal functioning, we hypothesized that Hypothesis 3) behavioral measures of emotion processing would correlate with the degree of self reported interpersonal problems. More specifically, we predicted that poorer performance on the emotion judgement task would be associated with greater interpersonal difficulty. Furthermore, we asked Empirical Question 3) Whether s ignificant relationships were accounted for by a particular interpersonal domain and Empirical Question 4 ) if these relationships differed between ADs and CCs. Summary of Aims In summary, the current study assessed the neurobehavioral correlates of emotion processing in alco hol dependent treatment seekers relative to non alcoholic controls. Using neurophysiological and behavioral measures of emotional facial expression processing, the investigation primarily aimed to contribute to this growing literature in two ways. First, we set out to replicat e previous research demonstrating atypical emotion processing in alcoholism, and to identify its neurophysiological
41 characteristics at earlier vs. later processing stages specificity to emotional content and select ivity according to emotion type Second, we intended to assess th e relationship between emotion processing and interpersonal functioning. Methods A description of the participants, experimental procedures, data acquisition and analysis is detailed below All procedures were approved by the University of Florida Health Science Center Institutional Review Board (protocol #125 2013). P articipants provided written informed consent prior to the collection of screening and testing data and were compensated for p articipating. Participants A total of 73 volunteers partici pated in the current investigation, including 34 alcohol dependent inpatient treatment seekers (AD; 28 men and 6 women) recruited from residential substance use treatment facilities and 39 healthy, non alcoholic community controls (CC; 15 men and 24 women) recruited via newspaper ads and flyers. Participants in b oth groups completed two screening sessions to determine eligibility. As part of the initial screening, part icipants completed pap er pencil questionnaires and cognitive assessments. Inventories concerned 1) demographics 2) negative affect, including assessments of depression (Beck Depression Inventory II (BDI II); Beck et al. 1996) and anxiety (State Anxiety Inventory (STAI); Spiel berger, 1983), 3) premorbid intellectual ability (Shipley Institute of Living Scale Verbal (SILS V); Zachary, 1986), 4) familial history of alcohol pro blems and 5 ) personal substance use histories, including brief assessments of chronicity, quantity and f requency of alcohol
42 use (Quantity Frequency Index (QFI); Cahalan et al., 1969) See Table 2 1 for further information regarding paper pencil screening measures and exclusionary cutoffs. Individuals who continued to qualify participated in the second scree ning session, during which vision was evaluated via Snellen chart (20/40 or better corrected vision was required to meet task demands), s elf reported medical histories (i.e. men tal/physical health conditions and current medications ) were obtained, and current/past psychiatric disorders were assessed Manual for Mental Disorders. DSM 5 instrumentation was not available at the start of investigation. Therefore, Axis I disorders were evaluated per DSM IV criteria using the computerized Diagnostic Interview Schedule IV (c DIS; Robins et al., 2000; American Psychiatric Association Task Force on DSM IV., 1994 ). Certain mental and physical health conditions were deemed exclusionary for both (AD and CC) groups These include d: 1) history of significant neurologic disorder/insult (e.g., stroke, epilepsy, coma, concussion), 2) medical conditions (e.g., HIV, Hepatitis C, chronic obstructive pulmonary disorder) or current use of medications (e.g., benzodiazepines, prescription op ioids) that can interfere with neurobehavioral function, 3) regular use of inhalants (e.g., ether), 4) lifetime diagnosis of a psychotic disorder or bipolar disorder, 5) current diagnosis of major depression (unrelated to substance use among ADs), and 6) s ignificant anxiety disorders, including lifetime diagnosis of obsessive compulsive disorder or agoraphobia and current diagnosis of panic disorder, social phobia, or post traumatic stress disorder. Given the prevalence of anxious and depressive symptomatol ogy among treatment populations, selection criteria related to these disorde rs were not further restricted in an effort to avoid recruiting a non representative
43 sample and to promote study feasibility. For similar reasons lifetime diagnosis of antisocial personality disorder and nicotine use / dependence were not exclusionary. Overall sample selection criteria were accompanied by additional group specific criteria. In line with previous investigations i npatient treatment seekers were required to meet DSM IV criteria for current alcohol dependence (determined via c DIS) be fully detoxified and have between 21 and 90 days of abstinence. This window of alcohol sobriety was chosen to limit the direct effects of detoxification, constrain variability that migh t be introduced by longer te rm abstinence, and ensure feasibility of participant recruitment within participating treatment facilities. A mong CC participants, current or lifetime DSM IV diagnosis of any substance dependence (excluding nicotine) was exclusi onary Furthermore, CCs were required to have a history of previous alcohol consumption but were not to exhibit drinking patterns that significantly exceeded 14 drinks per week or 4 drinks within a 2 hr period; U.S. Department of Agriculture and U.S. Department of Health and Human Services, 2010 ). It should be noted that DSM terminology and symptomatology for alcohol related diagnoses have undergone recent rev isions. Whereas the DSM IV distinguishes between alcohol abuse and dependence, the DSM 5 consists of a single disorder (Alcohol Use Disorder) which ranges in severity (mild, moderate, severe) dependent upon the number of criteria endorsed. Furthermore, a craving or strong desire to use alcohol has been added to these criteria in the DSM 5 and recurrent alcohol related legal problems have been removed. For the current study, DSM IV diagnoses were used for two reasons: 1) As noted previously, the DSM 5 diagn ostic instrument was not
44 available at the start of our investigation and 2) using DSM IV criteria allowed us to relate our findings to an existing literature. Importantly, measures of alcohol craving were collected ( Self Report Instrumen ts ) in an effort to generalize to the DSM V. Laboratory Protocol Participants were instructed to abstain from sedatives and sleep aids 24 hours prior to their scheduled testing session and to avoid the use of sedating allergy medications on testing day. I f participants typically consumed caffeine, they were advised to limit intake to one caffeinated beverage on the morning of testing. Transportation between the treatment facilities and our laboratory was provided for AD participants. Upon arrival, particip ants were queried about their adherence to pretesting guidelines and previously documented medication use was reviewed for potential changes. Additionally, a negative breath alcohol test, urine toxicology screen, and pregnancy test ( for women of childbeari ng potential ) were required for study participation. Participants who identified as current nicotine users were advised to continue typical use on the morning of testing to avoid potential withdrawal symptoms. The use of tobacco products is prohibited acro ss the University of Florida campus, where testing was conducted. Given the potential for withdrawal symptoms to confound interpretation of study outcomes, participants who did not typically abstain from nicotine for the length of testing were administered a 7 mg nicotine patch upon arrival (but no Ceballos et al., 2005; Nixon et al., 2007 ) suggest that this dose is well tolerated across sexes and does not differentially affect performance in ADs and CCs. The nicotine patch was removed by laboratory personnel after testing and participants were instructed to refrain
45 from nicotine/tobacco us e for several hours after its removal. As a precaution, blood pressure was measured periodically throughout the testing session. No adverse reactions were observed or reported. After completion of pre testing procedures, participants completed paper pencil inventories and computerized tasks. These measures are detailed below. Self Report Instruments Negative affect. The Beck Depression Inventory (BDI II; Beck, 1996) and State Anxiety Inventory (STAI; Spielberger, 1983) were re administered on the day of te sting to evaluate depressive and anxious symptomatology, respectively. The BDI II is a 21 item questionnaire for which summed scores reflect the severity of depressive symptoms experienced over the last two weeks. Levels of symptom severity and their corre sponding range of BDI II scores include: minimal (0 13); mild (14 19); moderate (20 28); severe (29 63). The STAI is a 20 item inventory for which participants report the intensity of their feelings at the present moment on a four point scale. The weighted scores for each item are summed and the total score is then age and sex corrected. Standardized scores range from 34, indicating low state anxiety, to 112, indicative of high state anxiety. Craving for alcohol. A lcohol craving was assessed among alcohol d ependent participants with the Desires for Alcohol Questionnaire (DAQ; Kramer et al., 2010; Love et al., 1998). The DAQ consists of 14 items that address the desire to drink, positive reinforcement, negative reinforcement, and the ability to control drinki ng behaviors. Participants were instructed to respond for their heaviest period of past drinking Scores range from 14 (no craving) to 70, with h igher DAQ scores correspond to higher levels of alcohol craving.
46 Interpersonal problems. The I nventory o f Inter personal Problems 64 (IIP 64; Horowitz et al., 2000 ) was used to assess Interpersonal difficulties. Participants respond ed to each of the 64 items on a five point Likert scale. This measure provides a total score (range: 34 99 ) and scores for 8 sub scales ( scoring range : 34 99 ), with higher scores indicative of greater interpersonal difficulty Sub scales address the following domains of interpersonal problems: 1) domineering/controlling: difficulty releasing control and a self identified tendency to be too controlling or manipulative; 2) vindictive/self centered: hostile dominance, anger, irritability, distrust of and suspiciousness toward others; 3) cold/distant: minimal feelings of affection for and connection with others; 4) socially inhibited: anxiety, t imidity, or embarrassment in social situations, as well as difficulty initiating social interactions and expressing feelings to others; 5) nonassertive: lack of self confidence and self esteem, difficulty dealing with social opposition, taking initiative, and being assertive; 6) overly accommodating: excessive submissiveness and difficulty denying the requests of others as well as self identification of being too gullible, exploitable, and easily taken advantage of; 7) self sacrificing: socially desirable traits (e.g., sympathy) have become excessive and problematic with the individual experiencing difficulty setting limits, maintaining boundaries, and putting their needs before the needs of others; 8) intrusive/needy: poor boundaries with other people, a strong need to feel engaged with and to hold the attention of others difficulty being alone and inappropriat e disclos ure of personal information or interject ion into matters that do not concern oneself. Overall and Sub scale scores were sex corrected.
47 Computerized Tasks To investigate emotional facial expression (EFE) processing, two face identification tasks, modeled after previously published electrophysiological paradigms (Maurage et al., 2008 b ), were administered: 1) an emotion judgement task (EJT) and 2) a sex judgement task (SJT) to ensure that EJT related findings were specific to emotion and observed impairment was not due to a more global visuospatial processing deficit. Given that general psychomotor slowing is occasionally observed in abstinen t alcoholics, a simple reaction time task (SRTT) was administered to control for group differences in psychomotor speed if necessary For all tasks, stimuli were presented at the center of a monitor and participants were instructed to respond via button pr ess with the index finger of their dominant hand. Response speed and accuracy were equally emphasized as important for successful task performance. All testing took place in a private, sound attenuated booth. Task descriptions are provided in the following sections and are outlined in the order in which they were administered (i.e. SRTT, SJT, EJT) Simple reaction time task. The SRTT was administered to assess simple psychomotor speed. For this task, an asterisks (*) was presented at varying intervals (1000 2500, and 5000 milliseconds (ms)) in a predetermined randomized order. The task consisted of 75 trials, with each stimulus presentation reflecting a single trial. Participants were instructed to respond via button press as soon as an asterisks appeared o n the screen. The stimulus remained on the screen until the participant responded, at which point the asterisks was replaced by a blank screen and the participant a waited the next stimulus presentation For each trial, the time interval
48 between asterisks o nset and participant response was recorded. The dependent variable of interest was mean reaction time (milliseconds) averaged across all trials. Sex judgement task and emotion judgement task. Face stimuli for both face identification tasks were derived us ing the Ekman stimulus set ( Ekman & Friesen, 1976 ), the most common set of emotional facial expression stimuli utilized in the emotional face processing literature Equal numbers of male and female posers, aged between 30 and 50 years, were selected in an effort to avoid sex and/or age related confounds. Task design and stimulus selection, detailed below, were guided by the work of Maurage and colleagues (2008 b ) For both face identification tasks, stimulus characteristics are reliably recognized and task difficulty is similar between emotion and sex judgement (Maurage et al., 2008b). For the SJT, eight ne utral expression faces (4 male and 4 female) were selected and male/female pairs were identified Morphing software (Fantamorph Version 5.4.3, Deluxe Ed ition) was applied to each male/female pair to create face stimuli that varied in the proportion of masculine/feminine features. Four morph levels were created for each pair, with the resulting stimuli depicting the following male/female proportions: 1) 5% / 95% (i.e. 95F) 2) 35% / 65% (i.e. 65F) 3) 65% / 35% (i.e. 65M) and 4) 95% / 5% (i.e. 95M). This procedure resulted in a total of 16 sex morphed face stimuli (4 pairs X 4 morph levels). See Figure 2 1 for an example of the stimuli obtained with these morphing procedures. To avoid participant expectation, sex morphed faces were presented in a pseudo random order. Randomization of stimulus presentation was achieved by dividing the SJT into 64 partitions, whereby each partition consisted of four trials (256 total trials) in which no single male/female pair or morph level was repeated.
49 These procedures resulted in a total of 64 stimuli per morph level. For each trial, a single face stimulus w as presented for 1500 ms, during which participants were to indicate whether the face appeared more male (left button) or more female (right button) using a response pad. Each stimulus was followed by a 300 ms interstimulus interval indicated by a fixation cross (+), which informed the participant that the response window for that trial had ended. This task took approximately ten minutes to complete. For the EJT, eight different posers (4 male and 4 female) were selected. In order to examine the effects of emotion type, task stimuli were developed using four facial expressions from each poser, including neutral, happy, angry, and sad expressions. Guided by the work of Maurage and colleagues ( 2008b ), investigation was limited to three emotional facial express ions (EFEs) to avoid participant fatigue and stimulus overexposure. To ensure adequate task difficulty and avoid ceiling effects, within poser morphing procedures were performed to create EFEs of varying intensities. For a given poser, each EFE was morphed with its respective neutral expression to represent 35%, 65%, and 95% of the target emotion (65%, 35%, and 5% neutral expression, respectively). These procedures resulted in 72 EFE stimuli (8 posers X 3 emotions X 3 morph levels). See Figure 2 2 for an ex ample of the EFE stimuli obtained using these morphing procedures. EFE stimuli were presented in a predetermined pseudo random order in a series of blocks. Within a single block, only two target emotions were presented and task demands required discriminat ion between these same two expressions. All possible two emotion combinations (i.e. happy angry, sad angry, happy sad) were each p resented in five blocks. Thus, t he EJT consisted of 15 blocks and 48 trials per block (8 posers X 2 emotions X 3 morph levels) resulting in a total of
50 720 trials (80 trials for any given emotion/morph combination) No single poser/emotion/morph stimulus combination was repeated within a given block. Prior to each block, participants were informed of the two emotions that were to be presented and their corresponding button assignment (left or right) on the response pad. For each trial, a single EFE stimulus was presented for 1500 ms, during which participants were to indicate which of two possible emotions was most prominently exp ressed by pressing the respective button. The EFE stimulus was followed by a 300 ms interstimulus interval indicated by a fixation cross (+), which informed the participant that the response window for that trial had ended. Button press assignment (L/R) fo r a given emotion alternated across task blocks and the order of block presentation was counterbalanced between participants and within group (AD vs. CC). The EJT took approximately twenty five minutes to complete. For both tasks, b ehavioral measures of i nterest were percent accuracy, reaction time (milliseconds) for accurate trials, and efficiency ratios (% accurate / average reaction time for accurate trials). Electrophysiology Recordings The electroencephalogram (EEG) was recorded during the SJT and EJT in a sound Limited, Morrisburg, Ontario). Participants were fitted with an elastic EEG cap (Electro Cap International, Eaton, OH) according to cranial circumference and nasion t o inion measurements, after completing the paper pencil assessments and the SRTT. Data were collected using a 64 electrode array mounted in an expanded International 10 20 System configuration, linked earlobe references, and a mid forehead ground. Electrod es were placed on the supraorbital and infraorbital perimeter of the left eye to
51 monitor eye blinks and movements. Electrode impedanc es were maintained below 10 kOhms with conductive gel, individually inserted into each electrode using blunted tip ped syrin ges. The continuous EEG was amplified at a gain of 10,000x and subjected to an analog 0.1 100 Hz band pass filter (Neuroscan 4.4 Acquire Compumedics USA, Charlotte, NC). Data were analog to digital converted at a sampling rate of 1000 Hz. During recordi ng sessions, participants were seated at a table with a mounted chin rest monitor using E Prime software (Psychology Software Tools, Inc., Sharpsburg, PA). Data Analysis A ll statistical analyses were performed using SAS Version 9.3 (SAS Institute, Adjustments for heteroscedasticity and multiple comparisons (Bonferroni post hoc corrections) were applied where appropriate. Participant Characteristics Data concerning demographics, negative affective, and substance use histories were analyzed for group differences, between alcohol dependent (AD) and community control (CC) participants using independent t tests a nd chi square analyses. Where appropriate, descriptive and dependent variables of interest were subject to Pearson correlations to identify potentially confounding relationships Behavioral Analyses SRTT performance was assessed for group differences using a t test. Separate generalized linear mixed models (GLMM) were used to analyze behavioral outcomes for the SJT and EJT. The use of GLMM, a particularly powerful and flexible model, allowed for the a pplication of data driven covariance structures and more accurate statistical
52 analysis when violations of normality and homoscedasticity were observed. Within these models, group (2; CC vs. AD) was designated as the between participants factor for both tas ks. For the SJT, morph level (4) was designated as the within participant factor. Two within participant factors were defined for the EJT, including emotion (3) and morph level (3). Although SJT and EJT within participant outcomes were not the focus of thi s investigation, within participant findings are reported in the results to provide a means of between study comparisons for general task related effects. ERPs EEG records were processed offline using EEGLAB Toolbox (Delorme & Makeig, 2004). Data were sub ject to a 30 Hz low pass digital filter and 24 dB attenuation Epochs were defined within the continuous record, starting 200 ms prior to face stimulus onset (baseline) and ending 1499 ms after it. Trials with incorrect responses were eliminated. Epochs th at contained gross artifacts (voltage recordings larger than 150 V) were rejected. Participants producing less than 4 0 percent of the total possible accepted epochs within a given block, for either the SJT (< 25 accepted epochs for any given morph level) or EJT (< 32 accepted epochs for any given emotion/morph combination), were excluded from task specific analys e s. After participant exclusion ( SJT: n= 9 3 CC and 6 AD; EJT: n= 12, 6 CC and 6 AD ), an average of 199.33 32.55 SJT and 542.17 88.58 EJT trials were deemed suitable for ERP component analysis. Remaining epochs were subjected to an independent component analysis (ICA; Jung et al., 2001) and an automated technique for identifying and removing artifacts (i.e. ADJUST; Mognon et al., 2011). For each participant, separate epochs were averaged and baseline corrected for the four morph levels of the SJT and for the nine emotion/morph combinations (e.g., happy expression at 65% intensity) of the EJT.
53 According to conventional ERP analysis proced ures ( Luck, 2014 ), N170 and P3 component coordinates, including measurement windows and electrode sites were determined via visual examination of group specific grand average waveforms and guided by the currently available literature. For both groups, P3 was measured between a post stimulus onset measurement window of 350 to 500 ms. Visual examination of the grand averaged waveforms suggested group dependent N170 latencies. Accordingly, measurement windows were defined separately for each group. N170 was m easured in post stimulus onset windows ranging from 110 to 180 ms for CCs and from 130 to 200 ms for ADs. Consistent with previous investigations (e.g., Maurage et al., 2008 b ) P3 was measured at PZ (midline electrode) and N170 was measured at T5 and T6 (c ontralateral temporal occipital sites). Given that hemispheric differences in component elicitation were not a primary focus of this investigation, and that the T5 and T6 sites yielded similar outcomes for dependent measures of interest, N170 measures were averaged across the T5 and T6 electrode sites to avoid unnecessary complication of data interpretation After measurement windows and sites were determined, individual mean amplitudes (i.e. average amplitude within a defined measurement window; Luck, 2014 ) and 50% area latencies (i.e. time from stimulus onset at which an epoch defined waveform reaches the midpoint of its calculated total area under the curve; Luck, 2014) were obtained (EEGLAB Toolbox) for the N170 and P3 components. These measures were cho sen over peak measures, as they are thought to be more reliable (e.g., less influenced by data filtering) and better suited to handle variability in noise level between groups (commonly observed in patient/control comparisons; for a more thorough review, s ee Luck, 2014)
54 Importantly, mean amplitude and fractional area latency, particularly those extracted from shorter duration components, are somewhat more sensitive to investigator defined measurement windows than peak measures (Luck, 2014). In an effort to ensure that the components of interest were accurately captured across participants, N170 and P3 amplitude were assessed for outliers (2 SD from the mean), separately, for each task and within group. When outliers were identified, their waveforms were ex interest occurred within the designated measurement window. When task and component specific latencies clearly occurred outside those defined via grand average waveforms, participants were excluded from the corresponding analysis. As a result, 3 and 6 participants were excluded from N170 analyses for the SJT and EJT, respectively. No participants were excluded from SJT or EJT P3 analyses. Although participant exclusion is less than ideal, th e documented benefits of mean amplitude and fractional area latency drove the decision to move forward with these measures in ERP analyses. N170 and P3 amplitudes, as well as P3 latencies, were individually assessed with separate GLMM analyses for the SJT and EJT. Similar to performance based analyses, group (2; CC vs. AD) was designated as the between participants factor for both tasks and within participant factors were defined as morph level (4) for the SJT, and emotion (3) and morph level (3) for the EJ T. As mentioned above, visual inspection of the grand average waveforms guided the use of different N170 measurement windows between groups. Therefore, N170 latency was analyzed within group; Separate analyses were conducted for the SJT and EJT using GLMMs that contained
55 within participant factors only. Similar to performance measures, within participant effects for ERP measures were not the focus of investigation, but are provided in the results for readers interested in general task related outcomes. Interpersonal Difficulties Overall IIP 64 scores (i.e. total scores) were investigated for group (AD vs. CC) differences with a one way ANOVA. M ixed effects ANOVA was used to test for group effects on IIP 64 subscale scores, with group (2) as the between p articipants factor and subscale scores (8 ) as within participant factors P otential relationships between interpersonal problems and behavioral EJT measures were assessed with Pearson correlations Correlation based a nalyses were initially conducted for to tal IIP scores and all nine emotion/morph combinations, within group. Where significant relationships were observed for either group, Pearson correlation coefficients were then calculated between the respective behavioral measure and the IIP subscales to d etermine if relationships could be accounted for by a particular interpersonal domain. Significant relationships observed among CCs and/or ADs, were also assessed for group specific correlation coefficients). General Considerations Acknowledging the potential for other factors to impact gr oup related outcomes variables that have previously been shown to influence emotion processing and/or general neurobehavioral function wer e given additional consideration. Although a ddressing all potential modulating factors is beyond the scope of this report, the current study focused on two variables sex and nicotine use.
56 Importantly, studies suggest that sex has the potential to modulate emotion processing in healthy populations (see Hamann & Canli, 2004 for a review) and influence alcohol related neurobehavioral compromise in various cognitive domains (see Nixon et al., 2014 for a review). Although a direct investigation of sex was not included as a covariate in SJT and EJT performance and ERP based analyses whenever possible, to control for its potential influence on study outcomes. Accordingly, a as a covariate. An other important factor with the potential to influence neurobehavioral function is nicotine use ( see Durazzo et al., 2010 for a review ). Given that the curre sample included nicotine users and non users, analyses were specifically conducted to determine whether or not nicotine use had an effect on the dependent variables of interest. Overall task performance as well as N170 and P3 amplitude and laten cy were individually assessed for potential differences between users and non users. Analyses of covariance (ANCOVAs) were conducted separately for the SJT and EJT with nicotine use status (current use vs. no use) designated as the between participants fac tor. Given the limited number of CC participants who endorsed current nicotine use, group (CC vs. AD) was included in the model as a covariate rather than an independent variable. Consistent with other SJT and EJT analyses, sex was added to the model as an additional covariate. Nicotine use status did not emerge as a significant factor in any SJT or EJT performance or ERP based analysis (all ps> 0.05) Therefore, this variable was not given any further consideration.
57 Table 2 1. Screening Measures and Excl usionary C utoffs a Beck, 1996, b Spielberger, 1983, c Zachary, 1986, d adapted from Mann et al., 1985; To calculate density, weighted scores are assigned to primary (1) and secondary (0.5) members with alcohol problems and divided by the total number of family members, e Cahalan et. al., 1969, f Exclusionary cutoffs apply to all participants unless otherwise noted. AD = alcohol dependent group. CC = community control group Figure 2 1. Sex Judgement Task: Illustration of Sex Morphed Stimuli. Figure depicts an example of the continuum from female to male face at the four mo rph levels presented in the Sex Judgement Task. Values indicate the proportion (%) of the dominant sex (F = female; M = male) depicted in the sex morphed stimulus. Face s timuli adapted by Neurocognitive Laboratory (University of Florida; PI: S.J. Nixon) from Ekman Stimulus S et ( Ekman & Friesen, 1976 ) Screening Measure Assessment Exclusionary Cutoff f Age < 25 or > 59 years Education < 10 or > 16 years Beck Depressi on Inventory (BDI II) a Depress ive s ymptomatology > 28 (i.e. severe symptomatology) State Anxiety Inventory (STAI) b State a nxiety Not exclusionary Shipley Institute of Living Scale Verbal (SILS V) c Verbal a bility < 13.1 Multi Generational Family Tree of Substance Use d Density of familial alcohol problems Not exclusionary Quantity Frequency Index (QFI) e Avg. oz. of absolute ethanol consumed per day in the 6 months prior to screening (CC) or treatment (AD) AD: QFI = 0 (indicates > 90 days abstinent) CC: Not exclusionary
58 Figure 2 2. Emotion Judgement Task: Illustration of Morphed Emotional Facial Expression Stimuli. Figure depicts an example of the EFE stimuli developed via morphing procedures and used in the Emotion Jud gement Task. This task consisted of three emotional expressions (happiness (top row), anger (middle row), and sadness (bottom row)), presented at three intensity levels. Values indicate the proportion of emotion depicted in each emotional/neutral expres sio n morphed stimulus (35, 65, and 95 percent intensity). Face stimuli adapted by Neurocognitive Laboratory (University of Florida; PI: S.J. Nixon), from Ekman Stimulus S et ( Ekman & Friesen, 1976 )
59 CHAPTER 3 RESULTS Descriptive Variables D emographic and affective measures are presented by group in Table 3 1 Overall, participants were between the ages of 25 to 59 years (M=42.41, SD=10.75) and had completed 11 to 16 years of education (M=14.14, SD=1.67). The majority of participants were Caucasian (78%). The sample also consisted of African Americans (18%) and day of laboratory testing, suggested relatively low levels of state anxiety (STAI; M=43.86, SD=8.82) and depressive symptomatology (BDI II; M=6.07, SD=6.51) across participants. When demographic and affective measures were compared between groups, significant differences emerged for educational achievement (t(71)=3.24, p=0.002) and depressive symptomatology (t(71)=3.38, p=0.001). Relative to AD participants, CCs had approximatel y one additional year of education and scored an average of 5 points lower on the BDI II. Consistent with inclusionary/exclusionary participant criteria mean BDI II scores for both groups indicated subclinical, minimal levels of depressive symptomatology. Given that these group differences were not particularly meaningful and neither education, nor the BDI II, significantly correlated with dependent measures of interest, these variables were not given further consideration. Analyses also revealed an unequa l distribution of male/female sex between CCs and ADs ( 2 =14.45, p= 0.0001 ). As discussed in the Methods, the data analysis strategy was designed to account for the potential influence of sex on dependent measures of interest. See
60 measures were equivalent between groups. Means and standard deviations for key alcohol related measures are presented for the AD group in Table 3 2 On average, ADs had completed two formal substanc e use treatments, reported a 15 year duration of alcohol use problems, and had a mean QFI reflecting pre treatment consumption of approximately 26 standard drinks per day. As expected, mean QFI among CCs (M=0.30, SD=0. 41 ; Range=0.00 2.21 ) was significantly lower (t(71)=8.63, p<0.0001), reflecting an average of less than one standard drink consumed per day. Analyses also indicated a higher density of familial alcohol problems among ADs (M density =0.35, SD=0.17 ; Range=0 .06 0.67 ), relative to CCs (M density =0.10 SD=0.11; Range=0.00 0.33; t(50)=6.26, p<0.0001 ). At the time of task administration, ADs had between 22 and 68 days of alcohol related abstinence. Regarding the use of other substances, 82% (n=28) of ADs endorsed the concurrent use of one or more drugs ( excluding nicotine) in the six months prior to treatment. The substances most commonly reported among concurrent drug users included marijuana (44%, n=15) and opioids (35%, n=12). Stimulant (cocaine: 18%, n=6; amphetamines: 12%, n=4), sedative (benzodiaz epines: 15%, n=5; muscle relaxers: 3%, n=1), and hallucinogen (3%, n=1) use was also reported. Four CC participants endorsed marijuana use in the six months prior to study participation. Consistent with inclusionary/exclusionary criteria, all participants tested negative for the presence of illicit substances on testing day. Current nicotine use was more prevalent among the AD group (76%, n=26) than the CC group (10%, n=4; 2 = 43.46 p< 0.0001 ). Importantly, planned analyses suggested no significant effect o f nicotine use on
61 neurobehavioral measures of interest. Consequently, this variable was not considered in subsequent analyses ( more information available in Neurobehavioral Measures of Task Performance Given the inves tigational focus on between group comparisons, within participant effects for the SJT and EJT (i.e. morph and emotion) are not reported here. The following sections detail group related neurobehavioral outcomes. Simple Reaction Time Task Analyses suggested equivalent SRTT performance between AD (M=481.10, SD=75.28 ms) and CC (M=482.80, SD=82.10 ms) groups (p>0.05). Therefore, it was not necessary to control for simple psychomotor abilities when assessing SJT or EJT performance. Sex Judgement Task Behavior. SJT performance measures are presented by group and morph level in Table 3 3 No significant main effects or interactions involving group were obse rved for any behavioral measure of interest. ERPs. ERP measures for the SJT are presented by gr oup and morph level in Table 3 4 No significant main effects or interactions involving group were observed for N170 amplitude or P3 amplitude or latency. Within group analysis of N170 latency suggested an effect of morph among CCs ( F(3, 31)=3.57, p=0.03) However, aft er correcting for multiple comparisons, posthoc analyses failed to reveal a significant difference between any morph level (ps>0.17). No significant findings emerged when N170 latency was analyzed within ADs.
62 Emotion Judgement Task Behavior. E JT performanc e measures are presented by group and emotion, within each morph level in Table 3 3 Group related main effect and interaction terms produced by o verall GLMM analysi s of accuracy, reaction time, and efficiency did not fully achieve statistical significanc e (ps>0.11). However, examination of group means suggested a potential difference in EJT accuracy that might have been concealed due to relatively small sample sizes and particularly complex statistical models. In an effort to address this suspected lack o f statistical power and adequately investigate the primary research questions of interest, EJT performance variables were reassessed with the most ambiguous morph level (35%) excluded from statistical models (i.e. follow up analysis with a 2 (group) X 3 (e motion) X 2 (morph) GLMM) Using this method, findings revealed a significant effect of group ( F(1, 70)=3.88, p=0.05) thereby comparisons suggested less accurate EFE identification among ADs (overall M=90.20, SD=5.73), relative to CCs (overall M=92.11, SD=4.18), across all emot ions and both morph levels ( Figure 3 1). Interestingly, no group by emotion, group by morph or group by emotion by morph interactions were revealed. However, examination of the GLMM estimated differences between group means suggested that the anger condition might have primarily driven the o bserved group main effect. Anger produced the largest mean difference ERPs. ERP measures for the EJT are presente d by group and emotion wit hin each morph level in Table 3 5 Analyses revealed a main effect of group on P3
63 amplitude (F(1, 59)=4.02, p=0.04 ; Figure 3 2 ); Across all emotions and morph levels, P3 amplitude elicited in response to EFEs was significantly sma ller among ADs ( overall M=5.56, SD=3.68 V), relative to CCs ( overall M=7.88, SD=3.64 V ). Grand average waveforms, measured at electrode PZ, are presented by group and emotion in Figure 3 3. Given that ERP measures are more sensitive than behavioral measu res in detecting subtle differences, it is not surprising that significant group related effects emerged sans reduction of the statistical model. In other words, it was not necessary to eliminate the most ambiguous morph level from analyses ; statistical po wer was adequate for detect ing significant group differences with these relatively sensitive neurophysiological measures. Examination of the GLMM estimated diffe rences between group means again suggested a more pronounced difference for anger. Although ang er did not produce the largest mean difference within all morph levels, the estimated group difference for anger at the 65% morph level was larger than the estimates of all other A65 H65 S65 =1.40 V; A9 5 H95 S95 A35 H35 =1.44 V; S35 =0.97 V). Regarding the other ERP measures, no significant main effects or interactions involving group were observed for N170 amplitude or P3 latency. Furthermore, within group a nalysis of N170 latency yielded similar outcomes for ADs and CCs; no significant effects emerged for emotion, morph, or their interaction in either group. N170 component characteristics collapsed across emotion and morph, are depicted in grand average waveforms at electrode site T6 in Figure 3 4.
64 Interpersonal Functioning One way ANOVA conducted for overall score on the IIP 64 revealed a significant grou p effect (F(1, 71)=13.17, p=0.000 5) ; H igher scores were observed among ADs relative to CCs. When IIP 64 subscale scores were assessed, mixed effects ANOVA yielded an overall effect of group (F(1, 71)=13.81 p=0.0004). Detailed characterization of this effect suggested significant group di fferences for all eight of the IIP 64 subscale than CCs. Means, standard deviations, and statistics regarding these data are p resented in Table 3 6 by group. In an effort to limit the number of statistical procedures conducted in the current investigation, analyses addressing the relationship between IIP 64 scores and behavioral measures of EJT performance were restricted to the investigation of accuracy. This performance m easure was selected over reaction time and efficiency given its relative sensitivity to group related effects. Correlational analyses conducted within group for overall IIP 64 scores and EJT accuracy revealed group dependent associations. No significant re lationships were observed among CCs for any particular emotion/morph combination. Conversely, correlational analyses conducted among ADs revealed significant relationships between the IIP 64 and measures of accuracy that were specific to anger. In particul ar, higher overall scores on the IIP 64 were associated with less accurate identification of angry expressions presented at 65% (r= 0.36, p=0.04) and 95% (r= 0.37, p=0.03) emotional intensity. When accuracy for these two EJT stimuli were subsequently asses sed for sub scale specific relationships among ADs, significant associations were specifically observed for three scales. Accuracy for angry EFEs at the 65% (r= 0.34, p=0.05) and 95% (r= 0.44, p=0.01) morph levels
65 intrusive/ needy scale of the IIP 64. Additional correlations were observed between accuracy at the 65% morph level and the scale (r= 0.37, p=0.03), as well as the 95% morph level and the 0.36, p=0.04). As e xpected, analyses performed among CCs yielded no significant correlations between angry EFEs and any of the IIP 64 sub scales. Comparison of group transformations revealed a significant difference be tween groups for the relationship between overall IIP 64 scores and accuracy for anger at the 65% morph level (CC: r=0.03; AD: r rep orted above; Z=1.63, p=0.05). C orrelation coefficients involving anger at the 95% morph level (CC: r= 0.18; AD: r reported a bove) did not significantly differ between ADs and CCs (p=0.19). With regard to the four sub scale correlations (reported above for ADs), statistically significant group differences were only observed for coefficients that characterized the relationship be intrusive/needy scale of the IIP 64 and accuracy for anger at the 65% morph level (CC: r=0.13; Z=1.96, p=0.03). All other group comparisons involving sub scale coefficients yielded trend level, albeit nonsignificant results (0.11> ps>0.07).
66 Table 3 1. Demographic and Affective Measures (By G roup) a Beck Depression Inventory II; values reported for testing day (Beck et al., 1996), b State Anxiety Inventory; values reported for testing day (Spielberger, 1983), c Significant effect of group. See te xt for more detail AD = alcohol dependent group. CC = community control group Measure CC (n=39) AD (n=34) M (SD) M (SD) Age (yrs.) 43.79 (12.12) 40.82 (8.84) Education (yrs.) c 14.69 (1.56) 13.5 (1.58) BDI II a c 3.74 (4.04) 8.74 (7.74) STAI b 42.03 (7.06) 45.97 (10.19) % (n) % (n) Race Caucasian 74 (29) 82 (28) African American 21 (8) 15 (5) Other 5 (2) 3 (1) Sex c Male 38 (15) 82 (28) Female 62 (24) 18 (6)
67 Table 3 2 Alcohol Related Measures (Alcohol Dependent P articipants ) a See Table 2 1 for information regarding derivation of this measure (adapted from Mann et al., 1985), b Quantity Frequency Index; 0.6 = 1 standard drink / day (adapted from Cahalan et al., 1969), c Desires for Alcohol Questionnaire ( Kramer et al., 2010; Love et al., 1998), d Pertains specifically to alcohol Measure M (SD) Range Total No. of Formal Substance Use Treatments 2.24 (1.33) 1 6 Chronicity of Alcohol problems (yrs.) 14.91 (10.90) 1 37 Density of Familial Alcohol Problems a 3.74 (4.04) 0.06 0.67 QFI b 15.69 (10.39) 0.37 45.82 DAQ c 42.03 (7.06) 22 69 Days of Sobriety on Testing Day d 41.00 (12.63) 22 68
68 T able 3 3. Sex J udgeme nt and Emotion J udgement : Task P erformance (By G roup) Table values shown are means (standard deviations) for accuracy and a CC: n=38; AD: n=33; Levels indicate the dominant sex portrayed in each sex morphed stimulus (M = male; F = female) and the degree (%) to which that sex was depicted (e.g., F95 = 95% female / 5% male), b CC: n=39; AD: n=34; Levels denote the emotion (H = h appy; A = angry; S = sad) and the intensity at which it was expressed in the respective face stimulus (e.g., H65 = 65% happy / 35% neutral expression), c Significant main effect of group for reduced model analysis of accuracy (i.e. excluding emotions prese nted at the 35% morph l evel). See text for more detail AD = alcohol dependent group. CC = community control group. Sex Judgement Task Morph a CC AD ACC (%) RT (ms) EFF ACC (%) RT (ms) EFF F95 91.39 (9.49) 775.01 (78.73) 1.1959 (0.2032) 92.84 (9.92) 774.95 (102.61) 1.2222 (0.2202) F65 83.79 (14.53) 814.21 (76.46) 1.0440 (0.2321) 85.97 (11.99) 813.87 (108.53) 1.0777 (0.2162) M95 96.47 (5.60) 748.61 (90.22) 1.3104 (0.1959) 97.43 (3.85) 765.65 (90.04) 1.2891 (0.1572) M65 84.22 (14.23) 855.78 (103.22) 1.0072 (0.2425) 84.07 (9.68) 879.20 (95.64) 0.9687 (0.1596) Emotion Judgement Task Emotion and Morph b c CC AD ACC (%) RT (ms) EFF ACC (%) RT (ms) EFF H35 84.09 (9.46) 854.14 (105.05) 1.0041 (0.1984) 84.96 (10.69) 832.66 (84.75) 1.0340 (0.1824) A35 65.39 (10.52) 897.25 (90.54) 0.7373 (0.1446) 64.92 (11.82) 890.06 (91.62) 0.7402 (0.1670) S35 83.69 (6.71) 889.14 (70.45) 0.9490 (0.1225) 82.57 (7.92) 876.46 (84.98) 0.9500 (0.1243) H65 97.07 (3.46) 746.66 (96.70) 1.3214 (0.1788) 95.56 (4.96) 732.16 (80.26) 1.3219 (0.1728) A65 89.44 (6.88) 853.08 (82.01) 1.0603 (0.1502) 85.12 (10.67) 845.85 (87.18) 1.0165 (0.1605) S65 87.95 (5.66) 829.06 (82.90) 1.0723 (0.1357) 87.59 (5.43) 820.73 (81.47) 1.0781 (0.1308) H95 96.70 (3.40) 718.60 (93.59) 1.3677 (0.1826) 95.96 (4.18) 711.02 (79.16) 1.3680 (0.1791) A95 92.98 (5.10) 816.74 (82.58) 1.1511 (0.1412) 90.21 (7.66) 805.92 (77.99) 1.1306 (0.1485) S95 88.33 (6.53) 819.18 (84.22) 1.0914 (0.1516) 87.37 (6.56) 809.09 (82.05) 1.0916 (0.1430)
69 Table 3 4 Sex Judgement Task : Electrophysiological R esults a Levels indicate the dominant sex portrayed in each sex morphed stimulus (M = male; F = female) and the degree (%) to which that sex was depicted (e.g., F95 = sex morphed stimuli depicting 95% female / 5% male), b N170 measurement windows differed between groups; comparison of group latencies is not approp riate. See text for more detail AD = alcohol dependent group. CC = community control group Morph Level a CC AD N170 (N=32) P3 (n=35) N170 (n=27) P3 (n=27) Amp (V) Lat b (ms) Amp (V) Lat (ms) Amp (V) Lat b (ms) Amp (V) Lat (ms) F95 0.61 (1.85) 147.59 (9.85) 5.73 (3.98) 434.69 (15.86) 1.20 (2.55) 163.78 (9.84) 4.00 (3.23) 429.67 (17.24) F65 0.65 (1.99) 144.39 (11.75) 5.62 (3.77) 431.83 (19.38) 0.76 (2.67) 163.94 (10.36) 3.22 (3.14) 428.07 (17.77) M95 0.54 (1.67) 144.14 (11.14) 5.52 (3.47) 431.26 (19.01) 1.39 (2.70) 163.24 (9.75) 3.79 (3.21) 431.26 (17.22) M65 0.69 (1.93) 143.25 (11.47) 4.80 (3.72) 426.00 (17.51) 1.14 (2.36) 162.33 (9.90) 3.30 (2.81) 425.96 (17.93)
70 Table 3 5 Emotion Judgement Task: Electrophysiological R esults a Levels denote the emotion (H = happy; A = angry; S = sad) and the intensity at which it was expressed in the respective face stimulus (e.g., H65 = stimulus depicting 65% happy and 35% neutral expression), b N170 measurement windows differed between groups ; comparison of group latencies is not appropriate, c Significant main effect of group on P3 ampl itude. See text for more detail AD = alcohol dependent group. CC = community control group Em otion and Morph a CC AD N170 (N=29) P3 (n=33) N170 (n=26) P3 (n=28) Amp (V) Lat b (ms) Amp c (V) Lat (ms) Amp (V) Lat b (ms) Amp c (V) Lat (ms) H35 1.54 (1.74) 144.45 (11.35) 7.54 (4.00) 431.61 (11.30) 1.08 (2.53) 164.60 (9.81) 5.06 (3.60) 428.82 (14.97) A35 1.60 (1.76) 146.67 (12.10) 7.17 (3.40) 429.73 (9.08) 1.18 (2.82) 165.79 (10.22) 5.18 (3.95) 427.14 (15.26) S35 1.41 (1.50) 144.60 (11.85) 6.74 (3.80) 430.73 (10.16) 1.40 (2.12) 164.87 (10.27) 4.74 (3.48) 429.17 (15.60) H65 1.29 (1.79) 144.60 (10.09) 8.19 (3.87) 430.45 (9.46) 1.52 (1.99) 163.88 (10.21) 6.20 (4.22) 428.50 (10.25) A65 1.66 (1.47) 146.84 (12.78) 8.07 (4.30) 431.18 (8.07) 1.48 (2.40) 164.33 (10.15) 5.38 (3.80) 432.54 (18.38) S65 1.50 (1.76) 147.05 (11.74) 7.69 (3.73) 431.06 (9.78) 1.47 (2.37) 163.94 (10.24) 5.24 (3.72) 429.18 (14.12) H95 1.61 (1.67) 144.14 (12.24) 9.12 (3.95) 430.73 (8.85) 1.34 (2.24) 161.42 (9.92) 6.50 (4.03) 433.54 (13.83) A95 1.49 (1.72) 147.84 (11.70) 8.55 (4.04) 433.36 (10.42) 1.66 (2.16) 165.44 (10.21) 6.21 (4.02) 430.11 (11.65) S95 1.70 (1.80) 145.03 (10.94) 7.88 (3.54) 432.18 (8.27) 1.47 (2.76) 163.73 (8.55) 5.52 (4.02) 430.32 (15.59)
71 Table 3 6 Inventory of Interpersonal Problems: Total and Subsc ale Scores (By G roup) Table values in center columns are group means (standard deviations) for IIP 64 (Horowitz et al., 2000) scores, a One way ANOVA and m ixed effects ANOVA revealed a significant main effect of group for total score and for all sub scale scores respectively detailed description of these sub scales. AD = alcohol dependent group. CC = community control group IIP 64 Scale CC AD Statistics a Total Score 50.64 (9.51) 59.62 (11.62) F(1, 71)=13.17, p=0.0005 Domineering / Controlling 49.51 (9.93) 59.74 (12.81) F(1, 71)=14.72, p=0.0003 Vindictive / Self Centered 49.13 (9.15) 57.41 (10.49) F(1, 71)=12.98, p=0.0006 Cold / Distant 51.21 (11.80) 56.97 (12.84) F(1, 71)=4.00, p=0.05 Socially Inhibited 51.05 (11.14) 56.59 (12.69) F(1, 71)=3.94, p=0.05 Nonassertive 51.82 (11.22) 57.82 (14.75) F(1, 71)=3.89, p=0.05 Overly Accommodating 49.46 (8.58) 57.85 (11.19) F(1, 71)=13.10, p=0.0005 Self Sacrificing 51.67 (8.60) 58.74 (11.13) F(1, 71)=9.35, p=0.003 Intrusive / Needy 50.46 (9.12) 58.32 (12.45) F(1, 71)=9.63, p=0.003
72 Figure 3 1 Emotion Judgement Task Accuracy : Group Main Effect. 2 ( group: AD vs. CC ) X 3 ( emotion: happy, angry, Sad ) X 2 ( morph level : 65%, 95% ) GLM M revealed less accurate performance among alcohol dependent participants, relative to community controls, across all emotions and morph levels (F(1, 70)=3.88, p=0.05). Error bars depict standard error.
73 Figure 3 2 Emotion Judgement Task P3 Amplitude : Group Main Effect. 2 ( group: AD vs. CC ) X 3 ( emotion: happy, angry, Sad) X 3 ( morph level : 35%, 65%, 95% ) GLM M revealed attenuated P3 amplitudes among alcohol dependent participants, relative to community controls, across all emotions and morph levels (F(1, 59)=4.02, p=0.04). E rror bars depict standard error.
74 Figure 3 3 Grand Average EEG Waveforms : Emotion Judgement Task ( Electrode PZ ) Y axis = mean amplitude. X axis = time from stimulus onset. Note the positive going deflection (P3) in both groups beginning at approximat ely 350 ms. Investigation of P3 Amplitude ( measured between 350 500 ms), revealed attenuated P3 amplitudes among alcohol dependent participants (AD) relative to community controls (CC) across all emotions and morph levels (F(1, 59)=4.02, p=0.04).
75 Figu re 3 4 Grand Average EEG Waveforms : Emotion Judgement Task ( Electrode T6 ) Y axis = mean amplitude X axis = time from stimulus onset. Note the negative going deflection (N170) in both groups around approximately 200 ms. N170 amplitude, as measured betw een 110 180 ms (CC) and 130 200 ms (AD ) did not significantly differ between groups. AD = alcohol dependent group. CC = community control group.
76 CHAPTER 4 DISCUSSION The current study was designed to contribute to a growing literature addressing alcohol related neurobehavioral concomitants of emotional facial expression (EFE) processing. More specifically, behavioral and neurophysiological measures were collected from alcohol dependent treatment seekers (AD) and non alcoholic community contr ols (CC) during emotional and non emotional task conditions. This i nvestigation focused on group comparisons and attempted to : 1) replicate previous observations of alcohol related E FE processing deficits and, if observed, 2) determine whether atypical neurobehavioral measures were specific to the processing of emotionally laden content or generalizable to the evaluation of other facial attributes 3) identify if atypical processing w as specific to negatively valenced emotions or to any one particular emotion, and 4) ascertain if EFE identification difficulties were associated with the ex tent of interpersonal problems. Behavioral i nvestigation of emotion processing with statistical ana lyses modeled to address a relatively small sample size yielded a significant difference between groups More specifically, the assessment of emotion judgement task performance with a two (group : AD, CC ) by three (emotion: happy, angry, sad) by two (morph: 65%, 95%) GLMM revealed less accurate emotion identification among ADs. These findings point to a subtle, but observable, alcohol related behavioral de ficiency that was generalizable across the positively and negatively valenced emotions examined in the current study ERP findings paralleled behavioral outcomes, with ADs exhibiting attenuated P3 amplitudes across all emotions, regardless of valence, emotion type, or EFE intensity. Importantly, the absence of group differences in behavioral and neu rophysiological
77 measures for the sex judgement task suggested that alcohol related neurobehavioral compromise was not generalizable to the explicit analysis of other facial attributes (i.e. masculine / feminine) and was, instead, specific to the analysis o f emotionally laden information. Some studies have also revealed equivalent behavioral outcomes using the emotions assessed in the current investigation and presenting them at similar intensities ( Philippot et al., 1999 ; Kornreich et al., 2003 2001 b ). Mor e specifically, alcoholics have previously exhibited less accurate EFE identification across happy, angry, and sad facial expressions. The study of anger and happiness, without sadness, has also yielded similar valence related findings (Maurage et al., 200 7 a ; Kornreich et al., 2001 a ). Furthermore, some of these investigations have extended the alcohol related behavioral deficit to include additional emotions, such as disgust ( e.g., Kornreich et al., 2001 b 2001 a ). Taken together, these findings, along with reinforce the hypothesis that the behavioral deficit is not valence specific. processing stages in these behavioral deficiencies. Investigation revealed typical N170 characteristics among ADs, but attenuated P3 amplitude in response to all EFE stimuli. Among the very few electrophysiological investigations of explicit EFE processing in recently abstinent alcoholics the neurobehavioral outcomes of one study correspond particularly well with the ones observed here. In this study, Maurage and colleagues (2008 b ) observed alcohol related behavioral deficits and altered P3 activity that was specific to emotion judgement and generalizable across happy, a ngry, and sad expressions. Investigation specifically revealed alcohol related reaction time delays as
78 well as attenuated P3 amplitudes in response to sad EFEs and delayed latencies in ier ERP components did not implicate the N170 as a specific contributor to the emotion recognition deficit exhibited by ADs. Although the exact nature of behavioral genera l pattern of results across our investigations seems to suggest alcohol related emotion identification difficulties that generalize to positively and negatively valenced emotions and to implicate later processing stages in these behavioral deficiencies. Investigation with other paradigms (e.g., oddball paradigm, multimodal visual conditions) has yielded comparable findings, providing evidence of EFE identification difficulties that are accompanied by atypical P3 elicitation, in the absence of N170 compone nt alterations (Maurage et al., 2008 d 2008 c ). Although these higher order perceptual processes are occasionally implic ated (e.g., Maurage et al., 2007b ), atypical activity at decisional processing stages appears to be more consistently observed. Important ly, visual inspection of the grand average waveforms in the current study guided N170 measurement in group specific epochs. Although amplitude did not differ between groups and results concerning latency were generally consistent within groups, circumstanc es did not permit a direct assessment of group effects on N170 latency. While conclusions canno t be made as to whether N170 latency significantly differed between ADs and CCs, grand average waveforms suggested that between group variability in N170 latency existed for both the emotion judgement and sex judgement task s (hence N170 measurement in group specific epochs for both tasks) Thus, if this study had permitted a direct comparison of N170 latency between groups, alcohol
79 related delays may have been observed for general face processing, as opposed to EFE processing specifically. Still, visual examination of the data only permits speculation. Therefore, further investigation is needed to characterize N170 latency in ADs, relative to CCs, und er emotional and non emotional facial expression processing conditions. Interestingly, some investiga tions have revealed an emotion processing deficit that is specific to or more pronounced for negatively valanced emotions; Findings have been noted f or both behavioral (e.g., Car mona Perera et al., 2014; Frigerio et al., 2002 ) and neurophysiological (e.g., Maurage et al., 2008 d 2008 c ) measures. Furthermore, a subset of these studies have suggested atypical EFE processing that is specific to a single n egatively valenced emotion (e.g., Frigerio et al., 2002 ; Maurage et al., 2008 c; Townshend & Duka, 2003 ) Although analyses conducted in the current study failed to reveal neurobehavioral impairment specific to any single emotion, the estimated differences between group means for accuracy and P3 amplitude were most produced for anger stimuli. Differences were particularly marked for moderate intensity presentations of anger. Our findings might point to a general neurobe havioral deficiency in emotion processi ng that exists for multiple expressions, but is more robust for certain emotions. Furthermore, these compromised functions might be most pronounced when the emotional expression is displayed at a lower, but decipherable, intensity. Indeed, previous researc h has suggested that abstinent alcoholics require greater emotional intensity in facial expressions for accurate identification ( Frigerio et al., 2002 ). Therefore, intensity seems to play a role in the magnitude of observable impairment. In fact, many of t he investigations that have observed an alcohol related deficit for
80 negatively valence emotions but preserved identification of positively valenced emotions have used face stimuli depicting maximum intensity emotional expressions (e.g., Car mona Perera et a l., 2014; Frigerio et al., 2002 ; Kornreich et al., 2013 ) If the emotion processing deficit is more pronounced for certain emotions, and in particular, negatively valenced emotions such as anger, observable impairment at maximum intensity presentations mig ht only persist for these emotions. Importantly, the current report is not the first to propose the enhanced deficit for anger processing Maurage and colleagues have previously emphasized the role of an ger in alcohol related emotion processing impairments (Maurage et al., 2008 d 2008 c ) and a recent meta analysis suggested that group related (AD vs CC) effects might be most pronounced for anger (although disgust wa s also noted; Bora & Zorlu, 2017 ). Given that the magnitude of observable impair ment appears t o vary by emotion and i ntensity of presentation also se ems to impact reported findings, a more dynamic approach to understanding alcohol related emotion processing deficits might be warranted. For some emotions, the deficit might emerge more readily, whereas other emotions may only reveal alcohol related compromise under specific conditions (e.g., lower intensity expressions, more demanding tasks, etc.). Rather than identifying discrete deficits that either do or do not definitivel y exist, emotion processing alterations among alcoholics might be better understood if approached with respect to the likelihood that an emotion might reveal impairment In other words, the deficit for a given emotion might belong on a continuum, rather th an in a category. Characterizing emotion processing in alcoholism has important implications for treatment and recovery outcomes. Facial expression is a chief component of non verbal
81 affective communication ( Eimer & Holmes, 2002 ). Deficient processing of emotional facial cues, and subsequent misidentification of emotions and intentions, might lead to negative social outcomes and interpersonal conflict (Uekermann & Daum, 2008). Interpersonal functioning has long been known to play a crucial role in recovery and maintenance of abstinence (Marlatt, 1996), and more recent research has demonstrated an association between poorer EFE processing and increased likelihood of treatment drop out (Foisy et al., 2007 a ). Although limited evidence exists, a direct associat ion between poorer emotion recognition and interpersonal problems has also been reported by two investigative teams (Kornreich et al., 2002; Maurage et al., 2009 ). Given the suggested import of EFE processing and interpersonal functioning in treatment outc omes, examining the relationship between the two might help provide a clearer understanding of emotion identification difficulties in recently abstinent alcoholics and better inform treatment efforts. As expected, the current investigation revealed higher levels of interpersonal difficulty among ADs, relative to CCs, in all interpersonal domains assessed. Of greater interest, we observed a relationship between EFE identification and interpersonal functi oning that showed group and emotion specificity. In particular, significant relationships were only observed among alcohol dependent participants and for angry facial expressions; Greater overall interpersonal difficulty was associated with less accurate i dentification of anger. Detailed characterization of this relationship further suggested that it might be attributable to certain interpersonal domains. Of the eight IIP 64 subscales, three exclusively shared a relationship with accuracy for anger, includi ng
82 higher scores indicated greater difficulty in these interpersonal domains and were associated with less accurate identification of anger among ADs. Subscales can be orga nized along two dimensions, affiliation and dominance. Interestingly, the three scales for which this relationship emerged fall within the affiliation dimension, with the erly speak to interpersonal problems that arise from a lack of self confidence, difficulty managing social opposition, vulnerability to exploitation and poor boundaries with others. suggested to reflect a loathing to express and feel anger. Perhap s these traits hinder alcoholic s ability to identify anger in others. In contrast, difficulty identifying ang er might promote uncertainty in the face of social opposition and exacerbate inappropriate responses to social cues. Although these data do not speak to causality, the relationships between these interpersonal domains and anger identification suggests that treatment efforts specifically aimed at ameliorating interpersonal problems within the affiliative dimension might also facilitate recovery of anger processing. Although this is not the first study to reveal a relationship between emotion processing and i nterpersonal functioning among recovering alcoholics, it is, to our knowledge, the first to investigate this relationship with respect to EFE identification accuracy and emotion type. M aurage and colleagues (2009 ) have touched on emotion specific associati ons in a previous study. However, their investigation concerned Likert scale ratings of emotional intensity for multiple emotionally laden stimuli, including facial expressions, body postures, voices, and short stories. Accordingly, correlational
83 analyses were conducted for emotional intensity ratings collapsed across all stimulus types. Nonetheless, their findings suggested a positive relationship between interpersonal problems and intensity ratings among ADs that specifically involved the anger scale. Alt hough methodological differences limit direct comparison with the current study the assessment of anger appears to share a distinct relationship with interpersonal difficulties. T aken together, the processing of anger might stand somewhat apart from other emotions. In the current investigation, a typical neural activity and misinterpretation of EFEs appeared to be more pronounced for anger, and poorer identification of angry expressions was associated with greater interpersonal difficulty. Although the alco hol related emotion processing deficit was more robust for, but not specific to, anger, the relationship it shared with interpersonal problems was. Given the previously noted role of emotion processing and interpersonal functioning in disease progression a nd treatment outcomes the role of anger in recovery from alcohol use disorder s merits further consideration. Importantly, the alcohol related behavioral deficit observed in the current investigation was subtle. However, this subtlety might be explained by the implementation of a two choice paradigm and investigation in a controlled laboratory setting. Emotion ide ntification deficits might be more pronounced in real world scenarios or with more complex task demands. However, previous studies have revealed a greater degree of behavioral impairment using similar methodological paradigms. One possible explanation for this discrepancy is variability in participant criteria for study participation. In an effort to address effects directly related to alcohol, the current
84 investigation enforced relatively conservative inclusionary/exclusionary criteria concerning comorbid mental / physical health conditions and current medication use. For example, a number of other investigations have recruited alcohol dependent samples without explicit exclusion for Hepatitis C, HIV, or ongoing benzodiazepine use (e. g., Philippot et al., 1 999 ; Maurage et al., 2008 b ). Some studies have also failed to detail exclusion on the basis of comorbid psychiatric disorders (e.g., Kornreich et al., 2002). Serious physical health ailments have the potential to influence neurobehavioral function (Heaton et al., 2004; Perry et al., 2008) and comorbid psychiatric conditions, such as major depressive disorder and social phobias, are shown to affect EFE processing (Joormann & Gotlib, 2006) Therefore, inclusion of participants with these characteristics might yield a more pronounced deficit. Our investiga tion suggests atypical emotion processing even when these factors are carefully controlled for, although, behavioral presentation of the deficit might be somewhat less pronounced. Interestingly, our investigat ion of behavioral efficiency did not reveal significant group differences for any task. This pattern might be explained by the observation that reaction times were slightly, albeit not significantly, faster among ADs relative to CCs. Given that accuracy am ong ADs was deficient but not severely impaired, a subtle tradeoff could have led to similar efficiency scores between groups. In other words, slightly faster responding at the cost of less accurate performance and in the absence of severe impairment, may have occluded a significant group difference in behavioral efficiency. To our knowledge, this is the first study of EFE processing in alcoholics to assess this measure. Therefore, additional investigation of behavioral efficiency
85 measures, particularly in larger sample sizes, will help to determine if a performance tradeoff exists for EFE processing Limitations Limitations for the current study should be noted. Investigation was limited to alcohol dependent inpatient treatment seekers without significant p sychiatric comorbidities or serious physical health ailments Although this facilitated the findings might not generalize to populations with more complex diagnoses. Oth er investigations have begun to characterize emotion processing in alcoholics with dual diagnoses (e.g., Steinmetz & Federspiel, 2012) and additional research is needed among sub populations presenting with and without comorbidities to determine whether or not various diagnoses modulate emo tion processing deficits. Furthermore, the and behavioral analyses had to accommodate for this issue in light of our complex design. Investigation with larger samples is ne cessary to confirm the findings reported here. Ongoing participant recruitment in our laboratory will allow us to address these issues and determine whether this pattern of results persists with an increased sample size. Relatedly, the current study did no t permit an investigation of sex main effects or interactions. Although we controlled for participant sex, sample characteristics restricted a meaningful analysis of its direct effect on alcohol related outcomes. Almost all of the previously published stud ies reviewed in this document ( Introduction) appear to have had similar constraints (but see Frigerio et al., 2002 and Fein et al., 2010 ), as alcohol dependent women are less frequently recruited than men. Given that sex has the potential to modulate e moti on processing in healthy populations (see Hamann & Canli,
86 2004 for a review) and influence alcohol related neurobehavioral compromise in various cognitive domains (see N ixon et al., 2014 for a review), future studies are encouraged to systematically investigate this factor when circumstances permit. Importantly, ongoing investigations in our laboratory are currently being conducted to address this understudied area. As noted previou sly, t he assessment of emotion processing with a two choice paradigm may have reduced our ability to reveal the full extent of impairment in this otherwise healthy alcohol dependent sample. The tasks used in the current study were designed to accommodate electro physiological assessment and limit potential complications during recording (e.g., excessive movement). Emotion judgement in realistic environments requires interpretation from a multitude of possible choices. Therefore, impairment might be more pronounced under real world circumstances. Additional investigation of similar alcohol dependent populations in more naturalistic settings and with more complex tasks are warranted to characterize the full extent of emotion processing deficits. In an effort to avoid participant fatigue, assessment of emotional facial expression identification was restricted to three emotions. Neurobehavioral assessment of emotion processing with tasks that incorporate additional emotions will provide further clarification regarding v alence and emotion specific outcomes. The investigation of disgust might be particularly informative. Similar to anger, disgust has been proposed as one of the emotions for which alcohol related compromise might be more pronounced. Future studies might al so benefit from assessing the relationship between interpersonal problems and performance measures for this emotion. Finally
87 electrophysiological assessment was limited to two ERP components. The examination of other ERP components reflecting different ps ychological processes will provide additional insight to the processing sta ges that contribute to emotion processing difficulties in abstinent alcoholics. Conclusions The current study demonstrated atypical emotional facial expression processing in alcoh ol dependent treatment seekers. Findings indicate subtle behavioral deficiencies and electrophysiological abnormalities that are specific to the evaluation of emotionally laden content, generalize to positively and negatively valanced emotions, and impli cate decisional processing stages in compromised emotion identification. These data also point to the enhanced role of anger in neurobehavioral compromise and in the relationship between emotion identification difficulties and interpersonal problems. This investigation addresses a relatively understudied topic with significant potential to foster a more comprehensive understanding of alcohol use disorders. Given th e dynamic interplay of emotion processing, interpersonal functioning, and treatment outcomes findings have important and clinically relevant implications. As the literature addressing this topic continues to grow, a clearer understanding of the deficit will emerge. Further investigation is needed to identify the emotions for which alcohol relate d compromise is most pronounced and to better understand the relationship between various emotions and specific domains of interpersonal difficulty. Although alcohol related compromise of emotion processing is far from being fully characterized, recent inv estigational efforts are elucidating the neurobehavioral and psychosocial components involved. These efforts, together with prospective studies, will advance our
88 for t he development of complementary therapies aimed at enhancing effective treatment and long term recovery.
89 LIST OF REFERENCES Acharya, R., & Dolan, M. (2012). Impact of antisocial and psychopathic traits on emotional facial expression recognition in alcoh ol abusers. Personality and Mental Health, 6 (2), 126 137. Adolphs, R. (2002). Neural systems for recognizing emotion. Current Opinion in N eurobiology 12 (2), 169 177. Adolphs, R., Damasio, H., Tranel, D., & Damasio, A. R. (1996). Cortical systems for the recognition of emotion in facial expressions. Journal of Neuroscience, 16 (23), 7678 7687. Agartz, I., Brag, S., Franck, J., Hammarberg, A., Okugawa, G., Svinhufvud, K., & Bergman, H. (2003b). MR volumetry during acute alco hol withdrawal and abstinence: A descriptive study. Alcohol and Alcoholism, 38 (1), 71 78. Agartz, I., Shoaf, S., Rawlings, R. R., Momenan, R., & Hommer, D. W. (2003 a ). CSF monoamine metabolites and MRI brain volumes in alcohol dependence. Psychiatry Research: Neuroimaging 122 (1), 21 35. American Psychiatric Association Task Force on DSM IV. (1994). Diagnostic and statistical manual of mental disorders: DSM IV (4th ed.). Washington, DC: American Psychiat ric Association. American Psychological Association. (2014). Understanding alcohol use disorders and their treatment. Retrieved from http://www.apa.org/helpcenter/alcohol disorders.aspx Bagga, D., Singh, N., Singh, S., Modi, S., Kumar, P., Bhattacharya, D. ... & Khushu, S. (2014). Assessment of abstract reasoning abilities in alcohol dependent subjects: A n fMRI study. Neuroradiology 56 (1), 69 77. Bates, M. E., Buckman, J. F., & Nguyen, T. T. (2013). A role for cognitive rehabilitation in increasing the ef fectiveness of treatment for alcohol use disorders. Neuropsychology R eview 23 (1), 27 47. Bates, M. E., Labouvie, E. W., & Voelbel, G. T. (2002 a ). Individual differences in latent neuropsychological abilities at addictions treatment entry. Psychology of Addictive Behaviors, 16 (1), 35. Batty, M., & Taylor, M. J. (2003). Early processing of the six basic facial emotional expressions. Cognitive Brain Research, 17 (3), 613 620. Beatty, W. W., Tivis, R., Stott, H. D., Nixon, S. J., & Parsons, O. A. (2000). Neur opsychological deficits in sober alcoholics: I nfluences of chronicity and recent alcohol consumption. Alcoholism: Clinical and Experimental Research, 24 (2), 149 154.
90 Beck, A T., Steer, R. A. & Brown, G. K. (1996). Beck Depression Inventory (2 nd ed.). San Antonio, TX: The Psychological Corporation. Bentin, S., Allison, T., Puce, A., Perez, E., & McCarthy, G. (1996). Electrophysiological studies of face pe rception in humans. Journal of Cognitive Neuroscience, 8 (6), 551 565. Blair, R. J. R., Morris, J. S., F rith, C. D., Perrett, D. I., & Dolan, R. J. (1999). Dissociable neural responses to facial expressions of sadness and anger. Brain, 122 (5), 883 893. Bora, E., & Zorlu, N. (2017 ). Social cogn ition in alcohol use disorder: A meta analysis. Addiction 112 (1), 40 48. Cahalan, D., Cisin, I. H., & Crossley, H. M. (1969). American drinking practices: A national study of drinking behavior and attitudes New Brunswick, NJ: Rutgers Center of Alcohol Studies. Campanella, S., Nol, X., & Tomberg, C. (2010). Cognitive e vent related potenti als and alcoholism. Journal of Psychophysiology, 24 (4), 231. Campanella, S., Petit, G., Maurage, P., Kornreich, C., Verbanck, P., & Nol, X. (2009). Chronic alcoholism: I nsights from neurophysiology. Clinical Neurophysiology, 39 (4), 191 207. Carmona Perera, M., Clark, L., Young, L., Prez Garca, M., & Verdejo Garca, A. (2014). Impaired decoding of fear and disgust predicts utilitarian moral judgment in alcohol dependent individuals. Alcoholism: Clinical and Experimental Research, 38 (1) 179 185. Ceballos, N. A., Tivis, R., Lawton Craddock, A., & Nixon, S. J. (2005). Visual spatial attention in alcoholics and illicit stimulant abusers: E ffects of nicotine replacement. Prog ress in Neuro P sychopharmacol ogy & Biol ogical Psychiatry, 29 (1), 97 107. Center for Behaviora l Health Statistics and Quality. (2015a). Results from the 2014 National Survey on Drug Use and Health: Detailed t ables. Retrieved from http://www.samhsa.gov/data/sites/default/files/NSDUH DetTabs2014/NSDUH DetTabs2014.htm#tab2 41b Center for Behavioral Health Statistics and Quality. (2015 b ). Behavioral health trends in the United States: Results from the 2014 National Survey on Drug Use and Health. HHS Publication No. SMA 15 4927, NSDUH Series H 50 Retrieved from http://www.sam hsa.gov/data/ Chanraud, S., Leroy, C., Martelli, C., Kostogianni, N., Delain, F., Aubin, H. J., ... & Martinot, J. L. (2009). Episodic memory in detoxified alcoholics: Contribution of grey matter microstructure alteration. PloS O ne 4 (8), e6786.
91 Chanraud, S., Martelli, C., Delain, F., Kostogianni, N., Douaud, G., Aubin, H. J., ... & Martinot, J. L. (2007). Brain morphometry and cognitive performance in detoxified alcohol dependents with preserved psychosocial functioning. Neuropsychopharmacology 32 (2), 429 438. Charlet, K., Schlagenhauf, F., Richter, A., Naundorf, K., Dornhof, L., Weinfurtner, C. E., ... & Gutwinski, S. (2014). Neural activation during processing of aversive faces predicts treatment outcome in alcoholism. Addiction biology 19 (3), 439 451. Clark, U. S., Oscar Berman, M., Shagrin, B., & Pencina, M. (2007). Alcoholism and judgments of affective stimuli. Neuropsychology 21 (3), 346. Coles, M. G. H., Gratton, G., & Fabiani, M. (1991). Event related brain potentials. In J. T. Cacioppo, & L. G. Ta ssinary (Eds.) Principles of psychophysiology: Physical, social and inferential elements (pp. 413 455). New York, NY: Cambridge University Press. Damasio, A. R. (1999). The feeling of what happens: Body and emotion in the making of consciousness Boston, MA: Houghton Mifflin Harcourt Dao Castellana, M. H., Samson, Y., Legault, F., Martinot, J. L., Au bin, H. J., Crouzel, C., ... & Syrota, A. (1998). Frontal dysfunction in neurologically norm al chronic alcoholic subjects: M etabolic and neuropsychol o gical findings. Psychological Medicine, 28 (5), 1039 1048. Delorme, A. & S. Makeig (2004). EEGLAB: A n open source toolbox for analysis of single trial EEG dynamics including independent component analysis. Journal of Neuroscience Methods, 134 9 21. Demirak ca, T., Ende, G., Kmmerer, N., Welzel Marquez, H., Hermann, D., Heinz, A., & Mann, K. (2011). Effects of alcoholism and continued abstinence on brain volumes in both genders. Alcoholism: Clinical and Experimental Research 35 (9), 1678 1685. D'Hondt, F., C ampanella, S., Kornreich, C., Philippot, P., & Maurage, P. (2014). Below and beyond the recognition of emotional facial expr essions in alcohol dependence: F rom basic perception to social cognition. Neuropsychiat ric Disease and Treatment 10 2177 2182. Don chi n, E., Karis, D., Bashore, T., Coles, M. & Gratton, G. (1986). Cognitive psychophysiology and human information processing. In M.G.H. Coles, E. Donchin, & S. Porges (Eds.) Psychophysiology: Systems, Processes, and Applications New York, NY: Guilford Press. Durazzo, T.C., Meyerhoff, D. J., & Nixon, S. J. (201 0). Chronic cigarette smoking: I mplications for neurocognition and brain neurobiology. International Journal of Environmental Research and Public Health, 7 3760 3791.
92 Eimer, M., & Holmes, A. (2002). An ERP study on the time course of emotional face processing. Neuror eport 13 (4), 427 431. Eimer, M., & Holmes, A. (2007). Event related brain potential correlates of emotional face processing. Neuropsychologia, 45 (1), 15 31 Ekman, P. & Friesen, W. V. (1976). Pictures of facial affect Palo Alto, CA: Consulting Psychologists Press. Fama, R., Pfefferbaum, A., & Sullivan, E. V. (2004). Perceptual learning in detoxified alcoholic men: contributions from explicit memory, executive function, and age. Alcoholism: Clinical and Experimental Research, 28 (11), 1657 1665. Fein, G., Bachman, L., Fisher, S., & Davenport, L. (1990). Cognitive impairments in abstinent alcoholics. Western Journal of Medicine, 152 (5), 531. Fein, G., & Cha ng, M. (2006). Visual P300s in l ong term abstinent chronic a lcoholics. Alcoholism: Clinical and Experimental Resear ch, 30 (12), 2000 2007. Fein, G., Key, K., & Szyma nski, M. D. (2010). ERP and RT delays in l ong term abstinent alcoholics in processing of emotional facial expressions during g e nder and emotion categorization t asks. Alcoholism: Clinical and Experimental Research, 34 (7), 1127 1139. Fein, G., Sclafani, V., Cardenas, V. A., Goldmann, H., Tolou Shams, M., & Meyerhoff, D. J. (2002). Cortical gray matter loss in t reatment nave a lcohol dependent i ndividuals. Alcoholism: Clinical and Experimental Research 26 (4) 558 564. Foisy, M L., Kornreich C. Fobe A. L. Pelc I. Hanak C. ... & Philippot, P. (2007a). Im paired emotional facial expression recognition in alcoh ol depend ence: D o these deficits persist with midterm abstinence? Alcoholism: Clinical and Experimental Research, 31 (3), 404 410. Foisy, M. L., Kornreich, C., Petiau, C., Parez, A., Hanak, C., Verbanck, P., ... & Philippot, P. (2007 b ). Impaired emotional facial express ion recognition in alcoholics: A re these deficits spe cific t o emotional cues?. Psychiatry Research, 150 (1), 33 41. Fried, I., MacDonald, K. A., & Wilson, C. L. (1997). Single neuron activity in human hippocampus and amygdala during recognition of faces and objects. Neuron, 18 (5), 753 765. Frigerio, E., Burt, D. M., Montagne, B., Murray, L. K., & Perrett, D. I. (2002). Facial affect percep tion in alcoholics. Psychiatry Research, 113 (1), 161 171. Galanter, M., Kleber, H. D., & Brady, K. (Eds.). (2014). The American Psychiatric Publishing textbook of substance a buse treatment (5 th ed.) Washington, DC: American Psychiatric Pub lishing
93 Gazdzinski, S., Durazzo, T. C., Mon, A., Yeh, P. H., & Meyerhoff, D. J. (2010). Cerebral white matter recovery in abstinent alcoholics a multimodality magnetic resonance study. Brai n, 133 (4), 1043 1053. Glenn, S. W., & Parsons, O. A. (1990). The role of time in neuropsychological performance: I nvestigation and application in an alcoholic population. The Clinical Neuropsychologist, 4 344 354. Glenn, S. W., & Parsons, O. A. (1992). Neuropsych ological efficiency measures in male and female alcoholics. Journal of Studies on Alcohol, 53 546 552. Hajcak, G., Weinberg, A., MacNamara, A., & Foti, D. (2012). ERPs and the study of emotion. In S. J Luck & S Kappenman (Eds.) The Oxford han dbook of event related potential components (pp. 441 474 ). New York, NY: Oxford University Press. Hamann, S., & Canli, T. (2004). Individual differences in emotion processing. Current Opinion in N eurobiology, 14 (2), 233 238. Hansenne, M. (2006). Event rela ted brain potentials in psychopathology: C linical and cognitive perspectives. Psychologica Belgica, 46 5 36. Hasin, D. (2003). Classification of alcohol use disorders. Alcohol Research and Health 27 (1), 5 17. Heaton, R. K., Marcotte, T. D., Mindt, M. R., Sadek, J., Moore, D. J., Bentley, H., ... & Grant, I. (2004). The impact of HIV associated neuropsychological impairment on everyday functioning. Journal of the International Neuropsychological Society, 10 (0 3), 317 331. Hermann, D., Smolka, M. N., Klein, S., Heinz, A., Mann, K., & Braus, D. F. (2007). Reduced fMRI activation of an occipital area in recently detoxified alcohol dependent patients in a visual and acoustic stimulation paradigm. Addiction B iology 12 (1), 117 121. Hoffman, L. A., Sklar, A. L., & Nixon, S. J. (2015). The effects of acute alcohol on psychomotor, set shifting, and working memory performance in older men and women. Alcohol, 49 (3), 185 191. Horowitz, L. M., Alden, L. E., Wiggins, J. S., & P incus, A. L. (2000). Manual for the inventory of interpersonal problems San Antonio, TX: Psychological Corporation. Joormann, J., & Gotlib, I. H. (2006). Is this happiness I see? Biases in the identification of emotional facial expressions in depression a nd social phobia. Journal of Abnormal Psychology, 115 (4), 705.
94 Julien, R. M., Advokat, C. D., & Comaty, J. E. (2010). Ethyl alcohol and the inhalants of abuse. In Primer of drug action (12th ed.) (pp. 431 478). New York, NY: Worth Publishers. Jung, T. P., Makeig, S., McKeown, M. J., Bell, A. J., Lee, T. W., & Se jnowski, T. J. (2001). Imaging brain dynamics using independent component a nalysis. Proceedings of the IEEE, 89 1107 1122. Jung, Y. C., Schulte, T., Mller Oehring, E. M., Namkoong, K., Pfefferbaum, A., & Sullivan, E. V. (2014). Compromised frontocerebellar circuitry contributes to nonplanning impulsivity in recovering alcoholics. Psychopharmacology 1 11. Kaplan, E. (1988). A process approach to neuropsycholo gical assessment. In T. Boll, & B. Bryant (Eds.), Clinical neuropsychology and brain function: Research, measurement, and practice. The master lecture series, Vol. 7 (pp. 1 27 167). Washington, DC: American Psychological Association. Knight, R. T. (1996). Contribution of human hippocampal region to novelty detection. Nature, 383 (6597), 256 259. Kopera, M., Jakubczyk, A., Suszek, H., Glass, J. M., Klimkiewicz, A., Wnorowska, A., ... & Wojnar, M. (2014). Relationship between emotional processing, drinking severity and relapse in adults treat ed for alcohol dependence in Poland. Alcohol and Alcoholism 50 (2), 173 179 Kornreich, C., Blairy, S., Philippot, P., Dan, B., Foisy, M. L., Hess, U., ... & Verbanck, P. (2001 a ). Impaired emotional facial expression recognition in alcoholism compared with obsessive compulsive disorder and normal controls. Psychiatry R esearch, 102 (3), 235 248. Kornreich, C., Blairy, S., Philippot, P ., Hess, U., Nol, X., Streel, E., ... & Verbanc k, P. (2001 b ). Deficits in recognition of emotional facial expression are still present in alcoholics after mid to lo ng term abstinence. Journal of Studies on Alcohol, 62 (4), 533 542. Kornreich, C., Brevers, D., Canivet, D., Ermer, E., Naranjo, C., Constan t, E., ... & Nol, X. (2013). Impaired processing of emotion in music, faces and voices supports a generalized emotional decoding deficit in alcoholism. Addiction, 108 (1), 80 88. Kornreich, C., Foisy, M. L., Philippot, P., Dan, B., Tecco, J., Noel, X., ... & Verbanck, P. (2003). Impaired emotional facial expression recognition in alcoholics, opiate dependence subjects, methadone maintained subjects and mixed alcohol opiate antecedents subjects compared with normal controls. Psychiatry R esearch, 119 (3), 251 260. Kornreich, C., Petit, G., Rolin, H., Ermer, E., Campanella, S., Verbanck, P., & Maurage, P. (2016). Decoding of nonverbal language in alcoholism: A perception or a labeling problem?. Psychology of Addictive Behaviors, 30 (2), 175.
95 Kornreich, C., Ph ilippot, P., Foisy, M. L., Blairy, S., Raynaud, E., Dan, B., ... & Verbanck, P. (2002). Impaired emotional facial expression recognition is associated with interpersonal problems in alcoholism. Alcohol and Alcoholism, 37 (4), 394 400. Kramer, J. R., Chan, G ., Hesselbrock, V. M., Kuperman, S., Bucholz, K. K., Edenberg, H. J., ... & Bierut, L. J. (2010). A principal components analysis of the abbreviated Desires for Alcohol Questionnaire (DAQ). Journal of Studies on A lcohol a nd D rugs, 71 (1), 150 155. Love, A., James, D., & Willner, P. (1998). A comparison of two alcohol craving questionnaires. Addiction, 93 (7), 1091 1102. Luck, S. J. (2014). An introduction to the event related potential technique Cambridge, MA: MIT press. Makris, N., Oscar Berman, M., Jaffin, S. K., Hodge, S. M., Kennedy, D. N., Caviness, V. S., ... & Harris, G. J. (2008). Decreased volume of the brain reward system in alcoholism. Biological psychiatry 64 (3), 192 202. Mann, K., Ackermann, K., Croissant, B., Mundle, G., Nakovics, H., & Diehl, A. (2005). Neuroimaging of gender differences in alcohol dependence: are women more vulnerable?. Alcoholism: Clinical and Experimental Research 29 (5), 896 901. Mann, R.E., Sobell, L. C., Sobell, M. B., & Pavan, D. (1985). Reliability of a family tree questionnaire for assessing family history of alcohol problems. Journal of Studies on Drug and Alcohol Dependence, 15 (1 2), 61 6 7 Marinkovic, K., Oscar & Harr is, G. J. (2009). Alcoholism and dampened temporal limbic activation to emotional faces. Alcoholism: Clinical and Experimental Research 33 (11), 1880 1892. Marlatt, G. A. (1996). Taxonomy of high risk s ituations for alcohol relapse: E volution and developme nt of a cognitive behavioral model Addiction, 91 ( S ), 37 49 Maurage, P., Campanella, S., Philippot, P., Charest, I., Martin, S., & de Timary, P. (2009). Impaired emotional facial expression decoding in alcoholism is also present for emotional prosody and body postures. Alcohol and Alcoholism, 44 (5), 476 485. Maurage, P., Campanella, S., Philippot, P., De Timary, P., Constant, E., Gauthier, S., ... & Verbanck, P. (2008 b ). Alcoholism leads to early perceptive alterations, independentl y of comorbid depressed state: A n ERP study. Clinical Neurophysiology, 38 (2), 83 97.
96 Maurage, P., Campanella, S., Philippot, P., Martin, S., & De Timary, P. (2008 a ). Face pro cessing in chronic alcoholism: A specific deficit for e motional features. Alcoholism: Clinical and Experimental Research, 32 (4), 600 606. Maurage, P., Campanella, S., Philippot, P., Pham, T. H., & Joassin, F. (2007 a ). The crossmodal facilitation effe ct is disrupted in alcoholism: A study with emotional stimuli. Alcohol and A lcoholism, 42 (6), 552 559. Ma urage, P., Campanella, S., Philippot, P., Vermeulen, N., Constant, E., Luminet, O., & de Timary, P. (2008 c ). Electrophysiological correlates of the disrupted processing of ange r in alcoholism. International Journal of Psychophysiology, 70 (1), 50 62. Maurag e, P., Joassin, F., Pesenti, M., Grandin, C., Heeren, A., Philippot, P., & de Timary, P. (2013). The neural network sustaining crossmodal integration is i mpaired in alcohol dependence: A n fMRI study. Cortex 49 (6), 1610 1626. Maurage, P., Philippot, P., Ve rbanck, P., Noel, X., Kornreich, C., Hanak, C., & Campanella, S. (2007 b ). Is the P300 deficit in alcoholism associated with early visual impairments (P100, N170)? An oddball paradigm. Clinical Neurophysiology, 118 (3), 633 644. Maurage, P., Rossignol, M., & Campanella, S. (2008 d ). The auditory visual integration of anger is impaired in alcoholism: An event related potentials study. International Journal of Psychophysiology, 69 (3), 221 222. Mognon, A., Jovicich, J., Bruzzone, L., & Buiatti, M. (2011). ADJUST: An automatic EEG artifact detector based on the joint use of spatial and temporal features. Psychophysiology, 48 229 240 Mller Oehring, E. M., Jung, Y. C., Sullivan, E. V., Hawkes, W. C., Pfefferbaum, A., & Schulte, T. (2013). Midbrain driven emotion and reward processing in alcoholism. Neuropsychopharmacology, 38 (10), 1844 1853. Namkoong, K., Lee, E., Lee, C. H., Lee, B. O., & An, S. K. (2004). Increased P3 amplitudes induced by alcoh ol related pictures in patients with alcohol dependence. Alcoholism: Clinical and Experimental Research, 28 (9), 1317 1323. National Institute on Alcohol Abuse and Alcoholism (2014). Measuring the burden of alcohol. Alcohol Alert, 87. Retrieved from http:/ /pubs.niaaa.nih.gov/publications/aa87/aa87.htm National Institute on Al cohol Abuse and Alcoholism. (2016 ). Alcohol use disorder: A comparison between DSM IV and DSM 5 NIH Publication No. 13 7 999 Retrieved from https://pubs.niaaa.nih.gov/publications/dsmf actsheet/dsmfact.pdf Nixon, S. J., & Bowlby, D. (1996). Evidence of a lcohol related efficiency deficits in an episodic learning t ask. Alcoholism: Clinical and Experimental Research, 20 (1), 21 24.
97 Nixon, S. J., & Glenn, S. W. (1995). Cognitive psychosocial perf ormance and recovery in female alcoholics. In M. Galanter (ed.), Recent developments in alcoholism volume 12: Women and alcoholism (pp. 287 307). New York, NY: Plenum Press Nixon, S. J ., Lawton Craddock, A., Tivis, R. D., & Ceballos, N. A. (2007). N effects on attentional efficiency in alcoholics. Alcoholism: Clinical and Experimental Research 31 (12) 2083 2091 Nixon, S. J., Prather, R. A., & Lewis, B. (2014). Se x differences in alcohol related neurobehavioral consequences. In A. Pfefferba um & E. V. Sullivan (Eds.), Alcohol and the nervous system, handbook of clinical neurology, 3rd series (pp. 253 272) Oxford, England: Elsevier. O'Daly, O. G., Trick, L., Scaife, J., Marshall, J., Ball, D., Phillips, M. L., ... & Duka, T. (2012). Withdrawal associated increases and decreases in functional neural connectivity associated with altered emotional regulation in alcoholism. Neuropsychopha rmacology 37 (10), 2267 2276. Oscar Berman, M., & M ffects on neurobehavioral functions and the brain. Neuropsychology R eview 17 (3), 239 257. Oscar Berman, M., Valmas, M. M., Sawyer, K. S., Ruiz, S. M., Luhar, R. B., & Gravi tz, Z. R (2014). Profiles of impaired, spared, and recovered neuropsychologic processes in alcoholism In A. Pfefferbaum & E. V. Sullivan (Eds.), Alcohol and the nervous system, handbook of clinical neurology, 3rd series (pp. 183 210) Oxford, England: El sevier. Parsons, O. A. (1987). Neuropsychological consequences of alcohol abuse: Many questions some answers. In O. A. Parsons, N. Butters, & P. E. Nathan (Eds.), Neuropsychology of alcoholism, implications for diagnosis and treatment (pp. 153 175). New York, NY: The Guilford Press. Parsons, O. A., Butters, N. E., & Nathan, P. E. (1987). Neuropsychology of alcoholism: Implications for diagnosis and treatment New York, NY: The Guilford Press. Parsons, O. A. & Leber, W. R. (1982). Alcohol, cognitive dysfun ction, and brain damage. In National Institute on Alcohol Abuse and Alcoholism: Biomedical processes & Consequences of alcohol use, Alcohol and Health Monograph 2 Rockville, MD: Department of Health and Human Services. Perry, W., Hilsabeck, R. C., & Hassa nein, T. I. (2008). Cognitive dysfunction in chronic he patitis C: A review. Digestive Diseases and Sciences, 53 (2), 307 321. Pfefferbaum, A., Adalsteinsson, E., & Sullivan, E. V. (2006). Dysmorphology and microstructural degradation of the corpus callosum: Interaction of age and alcoholism. Neurobiology of aging 27 (7), 994 1009.
98 Pfefferbaum, A., Desmond, J. E., Galloway, C., Me non, V., Glover, G. H., & Sullivan, E. V. (2001). Reorganization of frontal systems used by alcoholi cs for spatial working memory: A n fMRI study. Neuroimage, 14 (1), 7 20. Philippot, P., Kornreich, C., Blairy, S., Baert, I., Dulk, A. D., Bon, O. L., ... & V erbanck, Alcoholism: Clinical and Experimental Research, 23 (6), 1031 1038. Pitel, A. L., Eustache, F., & Beaunieux, H (2014). Component processes of memory in alcoholism: Patt ern of compromise and neural substrates In A. Pfefferbaum & E. V. Sullivan (Eds.), Alcohol and the nervous system, handbook of clinical neurology, 3rd series (pp. 211 225) Oxford, England: Elsevier. Polich, J. (2004). Clinical application of the P300 eve nt related brain potential. Physical Medicine and R ehabilitation C linics of North America, 15 (1), 133 161. Pol ich, J. (2007). Updating P300: A n integrative t heory of P3a and P3b. Clinical Neurophysiology, 118 (10), 2128 2148. Porjesz, B., & Begleiter, H. (1995). Event related potentials and cognitive function in alcoholism. Alcohol Health and Research World, 19 108 112. Quaglino, V., De Wever, E., & Maurage, P. (2015). Relations between cognitive abilities, drinking characteri stics, and emotional reco gnition in alcohol dependence: A preliminary exploration. Alcoholism: Clinical and Experimental Research 39 (10), 2032 2038. Robins, L. N., Cottler, L., Bucholz, K. K., Compton, W. North, C. S., & Rourke, K. M. (2000). The diagnos tic interview schedule, version IV St. Louis, MO: Washington University. Roopesh, B. N., Rangaswamy, M., Kamarajan, C., Chorlian, D. B., Pandey, A. K., & Porjesz, B. (2010). Reduced resource optimization in male alcoholics: N400 in a lexical decision para digm. Alcoholism: Clinical and Experimental Research, 34 (11), 1905 1914. Rossion, B., & Jacques, C. (2011). The N170: U nderstanding the time course of face perception in the human brain. In S. J Luck & S Kappenman (Eds.) The Oxford handbook of event rel ated potential components (pp. 115 142 ). New York, NY: Oxford University Press. Sacks, J. J., Gonzales, K. R., Bouchery, E. E., Tomedi, L. E., & Brewer, R. D. (2015). 2010 national and state costs of excessive alcohol consumption. American Journal of Preve ntive M edicine 49 (5), e73 e79. Salloum, J. B., Ramchandani, V. A., Bodurka, J., Rawlings, R., Momenan, R., George, D., & Hommer, D. W. (2007). Blunted rostral anterior cingulate response during a
99 simplified decoding task of negative emotional facial expressions in alcoholic patients. Alcoholism: Clinical and Experimental Research 31 (9), 1490 1504. Smith, M. E., & Oscar Berman, M. (1992). Resource limite d information processing in alcoholism. Journ al of Studies on Alcohol and Drugs, 53 (5), 514. Spielberger, C. D. (1983). Manual for State Trait Anxiety Inventory Palo Alto, CA: Consulting Psychologists Press. Stavro, K., Pelletier, J., & Potvin, S. (2013). Widespread and sustained cog nitive deficits in alcoholism: A meta analysis. Addiction B iology, 18 (2), 203 213. Steinmetz, J. P., & Federspiel, C. (2012). Alcohol related cognitive and affective impairments in a sample of long term care residents. GeroPsych: The Journal of Gerontopsychology and Geria tric Psychiatry, 25 (2), 83. Substance Abuse and Mental Health Services Administration. (2013). Results from the 2012 National Survey on Drug Use and Health: Su mmary of national f indings NSDUH Series H 46, HHS No. (SMA) 13 4795. Rockville, MD. Substance Abuse and Mental Health Services Administration ( 2014 ). National Surve y on Drug Use and Health Table 5.8A Substance dependence or abuse in the past year among persons aged 18 or older, by demographic characteristics: Numbers in thousands, 2013 and 2014 Retrieved from http://www.samhsa.gov/data/sites/default/files/NSDUH DetTabs2014/NSDUH DetTabs2014.htm#tab5 8a Sullivan, E. V., Rosenbloom, M. J., & Pfefferbaum, A. (2000). Pattern of motor and cognitive deficits in detoxified alcoholic men. Alcoholism: Cli nical and Experimental R esearch 24 (5), 611 621. Tipper, S.P. (2004). Selection for action: The role of inhibitory mechanisms. In D.A. Balota & E.J. Marsh (Eds.), Cognitive psychology: Key readings (pp. 224 232) New York, NY: Psychology Press. Tivis, R., Beatty, W. W., N ixon, S. J. & Parsons, O. A. (1995). Patterns of cognitiv e impairment among alcoholics: A re there subtypes?. Alcoholism: Clinical and Experimental Research, 19 (2), 496 500. Townshend, J. M., & Du ka, T. (2003). Mixed emotions: A recognition of specific emotional facial expressions. Neuropsychologia, 41 (7), 773 782. U.S. Department of Agriculture and U.S. Department of Health and Human Services. (2010). Dietary Guidelines for Americans, 2010, 7 th Editi on Washington, DC: U.S. Government Printing Office.
100 Uekermann, J., & Daum, I. (2008). S ocial cognition in alcoholism: A link to prefrontal cortex dysfunction?. Addiction 103 (5), 726 735. Uekermann, J., Daum, I., Schlebusch, P., & Trenckmann, U. (2005). P rocessing of affective stimuli in alcoholism. Cortex, 41 (2), 189 194. Wobrock, T., Falkai, P., Schneider Axmann, T., Frommann, N., Wlwer, W., & Gaebel, W. (2009). Effects of abstinence on brain morphology in alcoholism. European Archives of Psychiatry and Clinical N euroscience 259 (3), 143 150. World Health Organization. (2014). Global status report on alcohol and health 2014 Geneva, Switzerland: World Health Organization. Zachary, R. A. (1986). Shipley Inst itute of Living Scale: Revised m anual Los Angeles, CA: Western Psychological Services. Zywiak, W. H., Westerberg, V. S., Connors, G. J., & Maisto, S. A. (2003). Exploratory findings from the Reasons for Drinking Questionnaire. Journal of Substance Abuse T reatment 25 (4), 287 292.
101 BIOGRAPHICAL SKETCH Lauren A. Hoffman was born in Los Angeles, CA. She graduated from Reseda High School in 2007 and earned her Bachelor of Arts in psychology from San Diego State University in the spring of 2011. Later that year, Lauren entered the Behavioral and Cognitive Neuroscience doctoral program Psychology Department and joined the Neurocognitive Laboratory under the mentorship of Dr. Sara Jo Nixon Within this program, she was awarded a Master of Science degree in 2013 which concerned sex her doctoral research, Lauren was granted the opportunity to pursue her longstanding interest in a ddiction psychology Following receipt of her Doctor of Philosophy degree, Lauren will continue to pursue a career as an addictions researcher. She intends to refine her investigational skills and further develop her interests in substance use treatment by supplementing her neuroscience background with postdoctoral training in recove ry focused research. Ultimately, Lauren intends to conduct independent research that addresses treatment efficacy and methods of recovery aimed at reducing the risk of relapse.