The following projects are either ongoing or proposed. Not all projects on the list will necessarily be completed. This page is for general interest only.

Project 1: Cognition in Schizophrenia

Project 2: Cognition Associated with Caudate and Cerebellar Lesions

Project 3: Cognition and Huntington's Disease

Project 4: Cognition, Psychosis and Alzheimer's Disease

Project 5: Cognition and Mood Disorders in Brain Injury

References

PROJECT 1: COGNITION IN SCHIZOPHRENIA
Project Director: Nancy C. Andreasen, M.D., Ph.D.
Co-Investigators: Daniel O'Leary, Ph.D., Jane S. Paulsen, Ph.D., Jane Springer, Ph.D.
Funding Status and Proposed Duration: This project is partially funded by an RO1 grant to Dr. Andreasen entitled, "Phenomenology and Classification of Schizophrenia." Renewal is pending. The project will continue through 2003.
Specific Aims
1. To define basic, or fundamental, cognitive deficits in patients with schizophrenia and the relationship of these deficits to demographic, clinical, neuropsychological and outcome variables.
2. To further define the pattern of cognitive deficits in patients with schizophrenia and to explore the relationship among basic and complex cognitive abilities, clinical features, and outcome in patients with schizophrenia.
3. To further examine the course of cognitive impairment in patients with schizophrenia and to identify predictors of outcome.
4. To examine learning and memory patterns in patients with schizophrenia and to characterize clinical, premorbid, morphological and/or treatment factors with which they are associated.

Background and Rationale
There is presently a considerable amount of research investigating cognitive functioning in schizophrenia. Although most agree that cognitive performances are impaired in patients with schizophrenia (Heaton et al 1994), the overall degree and pattern of neuropsychological impairment varies greatly across patients (Goldstein 1994). Various strategies have been used to better understand cognitive heterogeneity in schizophrenia including differences in age of onset (Heaton et al 1994), subtype (Paulsen et al 1996), institutionalization or treatment effects (Green et al 1996; Mohamed et al in preparation), and variations in premorbid IQ (Palmer et al 1997; Nopoulos 1997a). Several investigators have stressed methodological issues as paramount in understanding cognition in schizophrenia (UCSD-CRC 3703, 1262, 4559). First, a major issue in neuropsychological research is the evaluation of generalized and specific cognitive deficits. Because schizophrenia patients perform poorly on many cognitive tasks, it is difficult to determine which performance deficits reflect dysfunction of discrete brain areas or systems and which are due to nonspecific, generalized factors. It has been suggested that future investigations of neuropsychological patterns could be evaluated by (a) statistical techniques for determining relative degrees of impairment across abilities (Chapman and Chapman 1973, UCSD-4245), (b) double dissociation UCSD 4265, and/or (c) qualitative analyses of response style and error patterns. Alternatively, some investigators argue that there exists a fundamental, or basic, cognitive deficit in schizophrenia whose presence can explain subsequent impairments on neuropsychological measures. Examples of this position include information processing deficits (Braff 1993; Nuechterlein et al 1994), willed action impairments (Frith et al 1991a; Frith 1992) or deficits in working memory (Goldman-Rakic, 1994). Most recently, we (Andreasen 1997a, b) described a unitary model of a fundamental cognitive abnormality called "cognitive dysmetria." One strategy of the Cognitive Neuroscience Research Unit is to begin to identify the cognitive deficits in schizophrenia exploring these and other models.

Although no known prospective longitudinal study has demonstrated a neurodegenerative course in schizophrenia, several recent papers have cited cross-sectional studies suggesting that cognition in schizophrenia worsens over time (Davidson and Haroutnian 1995). Preliminary analyses of data collected at the UCSD Center for the study of Psychosis in Late Life as well as preliminary analyses of the data collected at The University of Iowa suggest no cognitive deterioration over time (Gold et al submitted a; Nopoulos 1994). Given the potential significance of this question for treatment planning and understanding the etiology schizophrenia, longer follow up needs to be conducted to rule out the possibility of degeneration in older schizophrenia patients. The ongoing longitudinal work at Iowa seeks to provide outcome data to address these and several other issues in schizophrenia.

With the advent of newer treatments for schizophrenia, researchers have begun to address the clinical and treatment correlates of cognitive impairment. For instance, it is well established that anticholinergic medications interfere with various aspects of learning and memory performances in a dose-dependent fashion whereas neuroleptic medications are more likely to affect attention (positively) and motor functions (negatively). More recently, investigations have begun to examine how atypical antipsychotic drugs may impact cognitive functions differently than typical antipsychotic drugs. For instance, Clozapine has been reported to improve performance on measures of attention, verbal fluency, and verbal memory while transiently impairing verbal working memory (Hagger et al 1993; Lee et al 1994). Risperidone was found to improve verbal working memory and executive functions (Green et al 1997) and Olanzapine improved performance on measures of fluency, executive functions, reaction time, and verbal memory (McGurk and Meltzer, 1997). Our preliminary data with Olanzapine confirms these findings. Some recent research has even suggested that early and proper treatment of psychosis may predict better prognosis (Wyatt et al 1997). Given these recent findings, we consider it imperative to consider treatment variables in unison with cognitive performance in patients with schizophrenia.

Hypotheses
1. Performances on measures of cognitive dysmetria will be more impaired in schizophrenia patients with an earlier age of onset. Dependent variables include: Classical conditiong: acquisition rate, response latencey; Backward Masking: critical stimulus threshold, inter-stimulus interval; Flanker Test: facilitation score, interference score; Timing: time perception number correct, RTC accuracy; Procedural Learning: prism adaption average, motor tracking accurancy with feedback, motor tracking accuracy without feedback.
2. Performances on measures of cognitive dysmetria will be associated with measures of clinical neuropsychology that are sensitive to frontal lobe dysfunction. For instance, measures of cognitive dysmetria will be associated with speed of information processing, encoding and retrieval of information, organization and flexibility of thinking, but will NOT be associated with primary visuoconstruction, anomia or recognition memory.
3. Performances on measures of cognitive dysmetria will be associated with outcome in schizophrenia. For instance, poorer dysmetria will be associated with poorer outcome. Measures of cognitive dysmetria will be more closely related to outcome than more traditional measures of clinical neuropsychology.

Methods
Subjects: There are currently approximately 260 patients enrolled in our prospective longitudinal data base. This patient group consists of patients who are hospitalized for the first time and therefore are considered first episode as well as patients with a recent onset of psychosis (defined as individuals who have had five years or less of illness as dated from the time of initial hospitalization). Many of these patients are neuroleptic naive which minimizes the confounding effects of chronicity and effects of treatment which are areas of considerable concern for hypotheses about cognitive functions. Detailed description of the sample sizes, subgroups definition, subject inclusion and exclusion criteria, and sampling, recruitment and retention issues are contained in the Administrative Core, Assessment and Training Core, and the Diagnosis and Phenomenology Research Unit. Briefly, patients are included if they have a diagnosis within the schizophrenia spectrum. As the age of onset is often in adolescence we will recruit schizophrenia patients from the age of 12 upward. Although the male-female ratio in our original cohort of patients is around 2:5, with more males than females, we plan to make every effort to recruit a maximal number of female schizophrenia subjects since we are interested in exploring gender differences. The racial-ethnic composition of the subject population will reflect that of our catchment area. Racial ethnic minorities comprise approximately 5% of the Iowa population.

New patients are currently recruited at a rate of 80 each year. This new cohort of first episode or recent onset patients will be the subjects recruited for the new cognitive battery of tests.

Measures
The tests chosen for study in this project were selected based on one or more of the following criteria:
1. That the task has been implicated as sensitive to disruption in brain structure and/or function in one of the areas of interest (viz., prefrontal cortex, basal ganglia, thalamus, or cerebellum);
2. That the task had been previously emphasized in studies of cognition in schizophrenia;
3. That the task would allow a reasonable measure of basic cognitive processes and those emphasized by the construct "cognitive dysmetria," such that the face validity and prior work is consistent with timing, sequencing, or filtering dysfunction;
4. That the task was feasible to establish and collect data within the proposed time frame and in our patient samples.

Clinical Neuropsychology Tasks
A comprehensive battery is administered to each patient at intake and at the two, five, and nine-year follow-ups. This group of tests has been selected to offer the benefits of standardized tests as well as being useful in mapping functions of interest to schizophrenia. A full description of the tests is provided in the Assessment and Training core of the CRC renewal. Briefly, the battery consists of the WAIS-R, the Continuous Performance Task, Trails A and B, Circle A's, The Wisconsin Card Sorting Test, The Stroop Test, Controlled Oral Word Association Test, Category Fluency Test, Logical Memory (Wechsler Memory Scale), Rey Auditory Verbal Learning Test, Paired Associate Learning, the Rey-Osterreith Complex Figure, the Benton Visual Retention Test, the Finger Oscillation Test, and Purdue Pegboard.

Experimental Cognitive Tasks
In addition to the above standard group of tests, which have been given to CRC patients since it was established 10 years ago, we have developed a new set of experimental tasks designed to evaluate the construct of cognitive dysmetria. Five tasks are proposed: eyeblink conditioning, backward masking, facilitation and inhibition of motor responses, time perception, Responsive Temporal Consistency (RTC) and procedural learning and memory.

1) Eyeblink conditioning.
Specific Aim: To evaluate classical eyeblink conditioning in persons with schizophrenia. It is hypothesized that schizophrenics will display deficits in timing of the conditioned eyeblink response consistent with the cognitive dysmetria model of schizophrenia.
Rationale: Classical eyeblink conditioning is a motor learning task where subjects learn to produce a discrete motor response in the presence of a previously-neutral stimulus. Typically, eyeblink conditioning involves repeated pairings of an unconditioned stimulus (US), such as a mild air puff across the eye, and a neutral conditioned stimulus (CS), often a tone. With repeated pairings, the subject learns to blink when signaled by the CS and to produce a conditioned eyeblink response (CR) timed to occur maximally at the US onset. Computer control of stimuli presentation and eyeblink responses allows for millisecond temporal resolution of learning profiles for each subject. The classical eyeblink conditioning paradigm also allows for the separation of learning versus performance factors which is critical for interpreting neuropsychological studies of schizophrenia. Sensory or motor dysfunction in subjects, for example, can alter performance on learning tasks. In classical eyeblink conditioning nonassociative factors can be evaluated through control procedures, such as the presentation of US-alone trials to assess motor function. A final advantage of this paradigm for the study of schizophrenia is the extensive research on the neural systems involved in classical conditioning. Based on this research, the response patterns observed in schizophrenia can be interpreted relative to brain systems involved in various aspects of the task.

The neurobiological substrate of classical eyeblink conditioning has been extensively studied in humans and other species. An intact cerebellum is essential for acquisition of the conditioned response based on studies in the rabbit (Sears and Steinmetz 1990) and in humans (Solomon et al 1989). Information regarding the CS and US reach the cerebellum via mossy and climbing fibers, respectively, and appear to converge in the deep nuclei and cortex (reviewed in Kim and Thompson 1997). Research evidence suggests that output from the deep nuclei produce the learned motor response via the red nucleus (Chapman et al 1990). Measurement of the eyeblink response can also provide an indication of the site of cerebellar abnormalities. While lesions of the deep nuclei impair learning of the response, lesions of the cerebellar cortex can disrupt timing of the response (Mauk et al 1997). While the cerebellum is essential for learning to occur, other brain areas appear to be involved in certain aspects of conditioning. The hippocampus, for example, appears essential for processing complex stimuli during conditioning, such as for reversal learning in rabbits (Orr and Berger, 1985) and discriminative learning in humans (Daum et al 1991). Complex stimuli processing by the hippocampus, however, appears to require output from the cerebellum since lesions of the deep nuclei block learning-related activity in the hippocampus (Sears and Steinmetz, 1990).

The ability to evaluate brain function using the classical eyeblink conditioning paradigm has been exploited for a variety of patient populations. In addition to evaluating patients with traumatic brain lesions (Daum, et al 1991) classical conditioning has been used to study patients with mental retardation (Woodruff-Pak et al 1994) and amnestic syndromes (Gabrieli et al 1995). Of relevance for the proposed study, subjects with autism have been shown to exhibit rapid learning of the conditioned response but produce poorly timed responses (Sears et al 1994). Because of the overlap in autism and the negative symptoms in schizophrenics and based on the reported cerebellar abnormalities in autism (Bauman and Kemper, 1985) it is tempting to speculate that a similar pattern of cerebellar dysfunction occurs in both disorders. This possibility leads to the hypothesis that schizophrenics will exhibit rapid acquisition of conditioned responding but have poorly timed eyeblinks reflecting cognitive dysmetria.
Procedures: For classical eyeblink conditioning, subjects will be seated in a comfortable chair and view a silent movie. They will be informed that during the movie they will sometimes hear tones and also receive a mild puff of air across the eye. Eyeglasses for delivery of the air puff US and recording of the eyeblink will be placed on the subject (equipment purchased from San Diego Instruments). The air puff (1 psi) will be delivered through an air hose and eye closure will be recorded with an infrared photobeam. A tone CS (65 db, 1 kHZ) will be presented through earphones. For conditioning, subjects will be presented with 100 trials with the tone CS (450 ms) coterminating with an air puff US (50 ms) producing a 400 ms interstimulus interval. Intertrial interval will be randomly varied from 15 to 25 sec. Stimuli presentation and response recording will be controlled by computer.

Measures: Trials will be divided into 10-trial blocks for analysis. For each block, percent CR will be calculated based on the number of CR s (defined as eyeblink prior to air puff onset) within a 10-trial block. Timing of the CR will also be recorded based on the latency of the CR peak. Subjects will then be compared in acquisition rate and timing of the response. Conditioning procedures will produce several dependent measures. The number of trials to reach a criterion of 9 CR s out of 10 trials will be recorded as the measure of learning rate. The peak latency of the eyeblink will also be recorded to evaluate timing deficits in schizophrenia. Typically with learning, the eyeblink response will shift and an asymptotic level is obtained where subjects blink maximally at 400 ms after the tone CS where onset of the US occurs. Data will be transformed as appropriate to stabilize the variance. ANOVA will compare persons with schizophrenia and the comparison groups on acquisition rate and on differences in response latency after subjects reach learning criterion.

1) Backward Masking:
Specific Aim: To evaluate backward masking in persons with schizophrenia. It is hypothesized that schizophrenic patients will display deficits in the critical stimulus thresholds and critical inner stimulus intervals for backward masking and that performances will be associated with negative (and not positive) symptoms, other frontal and subcortical cognitive measures, structural and functional measures of brain, and indices of outcome.

Rationale: Schizophrenia patients have repeatedly demonstrated the inability to rapidly process information when tasks are timed or the processing load is relatively high. Schizophrenia patients show consistent deficits in the visual backward masking paradigm. In visual backward masking, an informational target stimulus is presented following an interstimulus interval by a masking stimulus that interferes with and/or interrupts target identification. Visual masking can be divided into an early component (e.g., up to 60 ms) that reflects the involvement of sensory-perceptual processes, and a later component that reflects susceptibility to attentional disengagement as the mask diverts processing away from the representation of the object. Backward masking performances have been shown to be reliable over time and associated with negative, rather than positive, symptoms in schizophrenia (Addington and Addington 1997). Backward masking has also been shown to be associated with frontal signs and soft neurological scores on a quantified neurological examination (Wong et al 1997). Although successful performance on the backward masking task requires visual identification and allocation of attention, one recent paper reported that performances in schizophrenia were associated with poor allocation of attention rather than perceptual or visual disturbances (Saccuzzo et al 1996). Recent research in unaffected siblings of persons with schizophrenia has shown that early, sensory-perceptual processes in backward masking deficits reflect enduring vulnerability to the disorder rather than only the symptoms of the illness (Green et al 1997). One recent paper suggested that visual backward masking was the only of several cognitive measures considered to be both sensitive and specific to schizophrenia (Suslow and Arolt 1997).
Procedures: We will replicate the procedure used by Braff (1993) which involves establishing critical stimulus thresholds and critical inner stimulus intervals for backward masking using an A versus a T as a stimulus. The Braff procedure requires differentiation of only a single visual feature. We will therefore also investigate critical stimulus durations and backward masking using more complex stimuli. A procedure will be used that establishes a critical stimulus threshold and backward masking interval independently for each visual hemifield (Casey 1991). Using this task, we have found that cognitive activation differentially changes backward masking functions into hemispheres. The task therefore provides an assessment of hemisphere-specific mechanisms involved in early visual processing. It is also relatively quick to administer (about 20 minutes) and has relatively low task demands.

2) Facilitation and Inhibition of Motor Responses.
Specific Aim: To evaluate facilitation and inhibition in patients with schizophrenia using measures of choice reaction time. It is hypothesized that schizophrenic patients will display deficits and that performances will be associated with negative (and not positive) symptoms, other frontal and subcortical cognitive measures, structural and functional measures of brain, and indices of outcome.
Rationale: Gray and colleagues (1991) and McKenna (1987) have proposed complex models of schizophrenia which integrate data from basic neuroscience, pharmacology, and psychiatry. These models propose dysfunctional activation within brain circuits resulting in cognitive abnormalities. The first model specifies that the system becomes overactivated by incoming sensory information and irrelevant motor sequences. That is, there is a breakdown in mechanisms which prevent insignificant information from being processed. As a result, schizophrenia patients are less able to inhibit or ignore stimuli in their environment and are less able to inhibit inappropriate behaviors. The second model posits that there is an error in the mechanism which integrates stored memories and ongoing behaviors. As a result, incoming perceptual information becomes erroneously labeled as important, when previous experience should mark the information as irrelevant. At a behavioral level, symptoms of schizophrenia such as delusions arise when an irrelevant stimulus is noted. A number of possible interpretations are generated, and an interpretation which previous experience should tell us is incorrect becomes selected. In the model proposed by McKenna, the initial bizarre interpretation of incoming perceptual information then derives abnormal selection and interpretation of future events. A decreased capacity to inhibit information has been noted by several schizophrenia researchers. It has been suggested that inhibition of distracting information plays a critical role in the types of behavior that are impaired in patients with schizophrenia, including remembering, concentrating, comprehending language, producing language, and creating and maintaining coherent streams of thought.

Active inhibition involves identification and suppression of irrelevant information so that limited attentional resources are not allocated to irrelevant stimuli. Inhibition occurs after selection of a target remains active for a long period of time and serves to prevent already-selected-against information from being activated. For optimal cognitive performance, selective attention and inhibition must work hand in hand. Inhibitory processes are likely involved in both early sensory processing (sensory gating) and later motor responses (response selection). To the extent that a failed inhibitory mechanism gives rise to both symptomatology of schizophrenia and the increased distractibility, changes in distractibility should be related to changes in symptomatology. Agents such as neuroleptics which correct the inhibitory deficit should be associated with reductions in the symptoms of schizophrenia and improvement of cognitive deficits.

The Flanker Task: This task is a choice reaction time paradigm in which a subject is required to respond when one of several possible target letters (e.g., an x or an o) appears in the center of a letter display. Reaction time is slowed when the central target letter is surrounded, or flanked, by letters that are potential targets but are not in the center of the display (e.g. an x flanked by o's). Slowed reaction time is a result of competition or interference between the response to the target and an automatic preparation of response to the flankers. In contrast, reaction time is faster when the central target letter is flanked by letters that require the same response as the center letter (e.g., an x flanked by x's). Facilitation is the result of redundancy of response information of the target and flanker stimuli. This phenomenon of increased reaction time on incompatible trials and decreased reaction time on compatible trials is referred to as the Flanker Compatibility Effect. This Flanker Compatibility Effect is robust to a number of experimental manipulations, including changes in distance between targets and flankers, and changes in size, color, and contrast of targets and flankers. In addition, the flanker task allows manipulation of levels of predictability. That is, some sets of trials were blocked and contained identical trial types. This allows subjects to form expectancies which can be used to predict the upcoming events. In contrast, other trial sets are random and contained many types of trials. Obviously, normal comparison subjects performed faster in the blocked sets than in the random sets. Authors have suggested that subjects are less influenced by the flanker letters in a blocked set because they are able to develop strategies to inhibit processing of flanker letters. In random sets, reaction time was affected by flanker letters because subjects could not predict what would occur on the next trial. In other words, knowing what will be required to inhibit makes it easier to do so. Models of schizophrenia that have been described earlier posit an inability of schizophrenia patients to accurately utilize previous knowledge or experience to guide present behavior. Schizophrenic subjects appear to be less affected by both compatible and incompatible flanker letters than control subjects (Jones et al 1991), and two studies (Elkins 1994; Kopp et al 1994) report that medicated subjects show less than normal facilitation on compatible trials than controls. The ability to initiate and/or inhibit motor responses is considered dependent upon the mesial orbital prefrontal cortex. Use of the Flanker Task is proposed in the current study to test the prediction of this model concerning the utilization of expectancies in schizophrenia. The effect of predictability will be tested by comparing reaction time in a predictable condition and reaction time in a random condition.

In summary, results from considerable previous research suggest that schizophrenia subjects suffer from an inability to suppress irrelevant information. Decrements in inhibitory mechanisms and/or increased distractibility reflect a state variable which worsens during acute episodes of psychosis and improves when symptoms remit, as well a trait variable observable in patients across the schizophrenia spectrum and in individuals at high genetic risk for the disease. It is likely, then, that compromise of inhibitory mechanisms underlies the cognitive dysfunction and symptomatology of schizophrenia. The proposed research paradigm is designed to explore the robustness and stability of inhibitory processing deficits in schizophrenia patients. The task can be used to evaluate the association between inhibitory processing, treatment effects, symptomatology, and other more traditional clinical neuropsychological tasks.
Procedures: On each trial, the subject will see a linear display of letters and is instructed to monitor the middle of the display for one of two targets (an x or an o). The subject is instructed to push one key on the keyboard when an x appears and another key on the keyboard when an o appears in the middle of the display. Displays remain on the screen until the subject has pressed either of the response keys or until 2500 milliseconds have elapsed. An intertrial interval of 500 milliseconds begins when the subject presses a response key or after 2500 milliseconds.
Predictable sets: 160 trials will be presented in 4 sets of 40 identical trials. Each condition described below will be presented for 1 block of 40 trials.
Random sets: 160 trials will be presented in 4 sets of 40 random trials. Each condition described below will be presented an equal number of times in each set of 40 trials. Four experimental conditions which are nested within the predictable and random sets are explained below:
No flankers. The target letter appears alone. This condition provides a baseline reaction time measure which will be used as a covariant in reaction time analyses to control for reaction time change that result from medications.
Neutral flankers. The target letter is flanked by letters which are not associated with any response. This condition serves as a control for the compatible and incompatible flanker conditions.
Compatible flankers. Flanker letters are associated with the same response as the target letter.
Incompatible flankers. Flanker letters are associated with a different response than the target letter. Reaction time is slowed when the central target letter is surrounded, or flanked, by letters that are potential targets but are not in the center of the display. Slowed reaction time is a result of competition, or interference, between the response to the target and automatic preparation of response to the flankers. In contrast, reaction time is faster when the central target letter is flanked by letters that require the same response as the center letter. For example, if the target letters are x and o, then response to an x flanked by o's will be slowed as a result of competition, or interference, between the x as the target and the o as the target. In addition, when the x is flanked by additional x's, facilitation will occur as a result of redundancy of response information of the target and flanker stimuli.
Measures: Each subject's median reaction time for each condition and trial type will be used in the analyses. There are several ways to analyze these data. First, to make our results comparable to those reported in previous studies, two different scores which describe the relative cost or benefit of flankers will be computed. The facilitation score (reaction time on compatible flanker trials minus reaction time on neutral flanker trials) reflects decreased reaction time in the presence of compatible flankers. The interference score (reaction time on neutral flanker trials minus the reaction time on incompatible flanker trials) reflects an increase in reaction time that results from the presence of incompatible flankers.

3) Time Perception and RTC:
Specific Aim: To evaluate time perception and RTC in schizophrenia. It is hypothesized that schizophrenics will display deficits in time perception and production and that performance will be associated with structural and functional measures of brain function.
Rationale: The neural substrates of several components of timing have been localized using pharmacological and lesion techniques in rats and, more recently, pharmacological and neuropsychological techniques in humans. Briefly, the evidence suggests that basal and frontal lobe structures and their interconnections are intimately involved in temporal processing in the seconds range. In addition, the cerebellum appears responsible for timing restricted to the milliseconds range. Some more complete findings are offered below. Animal research has demonstrated that lesions of the substantia nigra or the caudate putamen showed severe impairments in their ability to either generate or accumulate the pulses required to quantify the temporal duration of stimulus events. In addition, lesions to the frontal cortex and the nucleus basalis magnocellularis lead to overestimation of the expected time of reinforcement (Meck et al 1987a). Rats that received lesions of the frontal cortex or the nucleus basalis magnacellularis were able to time a single stimulus, but could not time two stimuli presented simultaneously. Methamphetamine administration to these rats caused immediate underestimation of duration, whereas haloperidol administration caused an immediate overestimation of duration. These findings have been interpreted as a change in the speed of the rat's internal clock, or pacemaker (Meck 1983, 1986, 1996). Physostygmine, which increases the effective level of acetylcholine in the brain, produced temporal underestimation, whereas atropine and scopolamine, which block acetylcholine receptors, produce overestimation (Meck and Church, 1987b).

Similar studies in humans have demonstrated that patients with Parkinson's disease make a higher percentage of underestimations and a significantly greater number of absolute errors. The administration of L-dopa significantly improved patients' performance on both the estimation and reproduction tasks. Patients with Alzheimer's disease showed normal accuracy but reduced precision emphasizing overestimation of duration. Patients with alcohol Korsakoff amnesia (damage to the hippocampal system) showed an underestimation of the target time. Finally, patients with olivopontocerebellar atrophy demonstrated increased timing variability. Most recently, a study of subjects at high risk for schizophrenia showed a larger difference between auditory and visual signal classification than did the normal control subjects. This was interpreted as a deficit in the timing mechanism in patients at risk for schizophrenia.

Timing is a critical component of any complex cognitive process, and dysfunctional timing of cognitive acts is hypothesized to be a central factor underlying cognitive dysmetria. We will explore timing functions using tasks described in Ivry and Keal (1989) who studied patients with lesions of different types on two measures of timing function. One task involved a production of timed intervals in which subjects try to maintain a simple rhythm. The other task measures a subject's ability to perceive small differences in the duration of two intervals. Ivry and Keal found that cerebellar lesion patients were impaired in both the production and perception of timing. The time perception and RTC paradigms will be adapted from Gibbon et al (1997) and Penney et al (1997), who developed computer-driven applications of this paradigm. Penny recently found that subjects at high risk for schizophrenia based upon family history showed a larger time perceptual deficit than normal comparison subjects.
Procedures:
(a) Responsive Temporal Consistency (RTC): A designated effector on a microswitch mounted on a wooden block is attached to the index finger of each subject's dominant hand. The subject is asked to press the microswitch which is connected to a computer which records all responses to the nearest millisecond. A series of 50 ms tones presented at regular intervals of 550 ms is presented to each subject and they are requested to tap at the same rate as the tones. After the subject's first response, 12 more tones are presented during which time the subject attempts to synchronize his or her responses. When the tones end, the subject is instructed to continue tapping at the same rate. After 31 self-paced taps have occurred, the computer signals the end of a trial. Feedback is provided indicating the mean interval produced with and without the tones (feedback) and the standard deviation of the inter-response intervals.
(b) Time Perception: Subjects compare successive intervals generated by two pairs of tones. Each 1000 Hz tone is 50 ms in duration and played at a volume of 73 dB. The stimulus onset asynchrony between the first pair of tones is 400 ms and was designated the standard interval. One second after the offset of the first pair, the second pair is presented. On half the trials the interval between the second pair of tones was chosen in order to estimate the lower threshold (i.e., the point at which the subject would respond "shorter" on 90% of the trials) and on the other half of the trials the upper threshold was sampled (i.e., the point at which the subject would correctly respond "longer" on 90% of the trials). The task was conducted on a Macintosh computer and separate response keys are labeled "longer" or "shorter." This task can be made more complex with the addition of a second task, as suggested by Gibbon and Penney.
Measures:
RTC mean interval produced with feedback and RTC mean interval produced without feedback are obtained. Number correctly identified on time perception is obtained.
1) Procedural Learning and Memory:
Specific Aim: To evaluate the acquisition and storage of motor information in patients with schizophrenia. It is hypothesized that patients with schizophrenia will all demonstrate impairments in the acquisition of motor information, but that only patients with tardive dyskinesia will also reveal impairments in the retention of information.
Rationale: Numerous studies have demonstrated learning and/or memory impairments in schizophrenia patients (Calev et al 1983; Goldberg et al 1989; Saykin et al 1991; Gold et al 1992; Schwartz et al 1992; Spitzer 1993). Evidence for learning and memory abnormalities has been recorded for different types of stimuli (words, sentences, stories, nonsense syllables, figures) in a variety of research paradigms (paired associates, free recall, cued recall, recognition formats, repetition priming), and within a large range of schizophrenic patient groups (acute, chronic, process, reactive). Most investigators agree that learning and memory impairments are one of the most striking and consistent neuropsychological findings in schizophrenia. There is little consensus, however, on the profile of spared and impaired learning and memory components in schizophrenia. Recent reports have suggested that the pattern of memory deficits in schizophrenia does not readily conform to profiles typically seen in amnesia (Goldberg et al 1989), temporal lobe epilepsy (Gold et al 1994), or subcortical and cortical dementias (Paulsen et al 1995). Some investigators have suggested that methodological differences are responsible for heterogeneous findings in the research on memory functions in schizophrenia (Levin et al 1989; Heinrichs 1993). Several studies have found that specific disease factors (e.g., severity and type of psychopathology), treatment factors (i.e., neuroleptic and anticholinergic medications), subject factors (e.g., age of onset, age, duration of illness), and measurement factors (i.e., type of memory assessed) are associated with learning and memory performance (Koh and Peterson 1978; Calev et al 1983; Spohn and Strauss 1989; Cassens et al 1990; Gold et al 1991; Eitan et al 1992). In recent years, researchers have begun to address the clinical and treatment correlates of cognitive impairment in schizophrenia. For instance, several studies have reported that schizophrenic patients with more prominent negative symptoms perform worse than those with more prominent positive symptoms on cognitive examination (Owens and Johnstone, 1980; Green and Walker, 1985; Breier et al 1991; Perlick et al 1992). In addition, there exists some consensus that anticholinergic medications interfere with various aspects of learning and memory performances in a dose-dependent fashion whereas neuroleptic medications are more likely to affect attention (positively) and motor functions (negatively)(Medalia et al 1988; Spohn and Strauss, 1989; Cassens et al 1990; Gold et al 1991, 1992; Bilder et al 1992; Goldberg et al 1993; Heaton et al 1994). Much less agreement has been established, however, on specific subject and measurement factors affecting learning and memory processes in schizophrenia.

Over the past few decades multiple memory systems have been described in the human brain (Squire and Zola-Morgan 1988; Tulving and Schacter 1990). Explicit or declarative memory refers to conscious recall of events or facts, whereas implicit or nondeclarative memory does not involve conscious effort and includes classical conditioning, priming, and skill or habit formation. Explicit memory can be further subdivided as being either semantic or episodic (Tulving 1972). Semantic memory refers to storing factual information about the world, whereas episodic memory deals with recollecting a personally experienced event. Many studies have demonstrated that the operation of explicit memory depends on the integrity of a corticolimbic subsystem comprised of the hippocampus, parahippocampal gyrus, and amygdala. The concept of explicit memory is supported by more than 30 years of research with lesioned monkeys (Mishkin 1978; Zola-Morgan and Squire 1985; 1986) and human patients (Salmon et al 1992). In contrast, some components of the implicit memory system have been ascribed to the basal ganglia. For instance, skill learning (i.e., procedural learning) is deficient in animals with caudate lesions and humans with progressive deterioration of the basal ganglia. Although several investigators have reported memory deficits in schizophrenia, few have described memory impairments according to the set of subsystems that have been established in the current literature on memory. Even fewer researchers have related their behavioral findings to neural systems associated with memory. Interestingly, the temporolimbic, frontal, and basal ganglia regions critical for different forms of memory are also implicated in the etiology of schizophrenia (Goldman-Rakic 1990). Although there exists a significant amount of research on explicit memory functions in patients with schizophrenia there is little consensus regarding the neuroanatomical specificity (i.e., temporolimbic or dorsolateral prefrontal cortex) of the learning and memory deficits. Few investigators have examined implicit memory functions in schizophrenia, and the studies that have been conducted have revealed contrasting findings. For instance, Schwartz et al (1992) reported a dissociation among implicit learning (impaired skill learning and intact priming) whereas others failed to find implicit memory impairments (e.g., Goldberg et al 1993; McKenna et al 1995; Schmand et al 1992). A number of research studies support the notion that skill-based learning and skillful motor performance depend on the integrity of several basal ganglia structures and cortico-striatal circuits (Heindel et al 1991; Paulsen et al 1993). The dorsolateral prefrontal cortical circuit in particular, involving the caudate, putamen, and supplementary motor and prefrontal cortex, appears to mediate the development of motor programs and permit the organization of motor sequences prior to their execution.
Procedures: We will explore the initiation and maintenance of motor action using tasks previously used in both animal (Bossom 1965) and human studies (Malenka et al 1982; Paulsen et al 1993). The first experiment examines visual adaptation to laterally displaced vision during trials with and without visual feedback from motor responses. The Prism Adaptation Task offers strengths over previously used measures in that it has a measure of baseline motor control from which the effects of feedback can be estimated above and beyond primary motor dysfunction. Previous work has suggested that Prism Adaptation performance is impaired in patients with basal ganglia and cerebellar damage. Each subject is instructed to mark the position of a target line with the middle finger while wearing goggles that distort the vision. The main unit of analysis will be post-adaptation average error minus baseline average error. The experimental procedure for a session is as follows:
Baseline: 6 practice trials without the prism goggles and no feedback.
Pre-adaptation: 12 test trials with prism goggles and no feedback.
Adaptation: 30 trials with goggles and feedback.
Post-adaptation: 12 test trials with goggles and no feedback.
Aftereffects: 12 test trials with the goggles removed and feedback.
The second task involves the measurement of reaction time, velocity, and accuracy during movements toward a moving target. The Tracking Task has been found to be sensitive to basal ganglia and cerebellar dysfunction. The main advantage of the Tracking Task over other measures is precise measurement of movement parameters.
Subjects are seated before a computer, given a joystick, and told to try to put a small dot in the center of a circle as quickly as possible when it jumps upwards. The protocol is as follows:
Baseline: 4 jumps at each of 5 amplitudes are presented in random order. This baseline reaction time is used to examine baseline differences among groups.
Learning with constrained stimulus: 5 groups of 8 jumps at 1 amplitude.
Learning with non-constrained stimulus: 5 groups of 8 jumps in sequence will be followed by 1 block presented in random order to allow measures of sequence-specific learning and reaction-time task learning.
Feedback control: 5 blocks of the same 8-item sequence are presented where the target marker disappears immediately after the movement is begun, followed by 1 block presented in random order.

Both of the preceding tasks, Prism Adaptation and the Tracking Task, have the ability to evaluate the utility of the schizophrenia patient to use feedback in perfecting cognitive performances. Based upon prior research, varying performance patterns are anticipated based on the neuroanatomical dysfunction. For example, patients with a prefrontal or basal ganglia dysfunction may demonstrate slow but accurate adaptation with impaired post-adaptation. That is, these patients will be able to acquire accurate responses but will be unable to retain these motor responses. Alternatively, patients with cerebellar dysfunction will be unable to both acquire and maintain the performance regardless of whether feedback is provided.

Other Measures
A comprehensive evaluation is conducted on each subject enrolled in the study and the evaluation is described in detail in the Assessment and Training Core. Briefly, the evaluation consists of the following: Comprehensive Assessment of Symptoms and History (CASH), Psychosocial Symptoms You Currently Have (PSYCH base), Scale to Assess Unawareness of Mental Disorder (SUMD), Neurological Examination, Childhood Standardized Test Scores (i.e., ITBS), Childhood home movies, Cognitive Battery, MRI Scan, videotaped interview highlighting symptoms, Birth and Developmental History (describing proband), Family History Research Diagnostic Criteria (FH-RDC) (proband's family).

Data Analysis
Bivariate and multivariate correlation will be used to address Hypothesis 1 which speculates a positive correlation between age of onset and performance on our measures of cognitive dysmetria. Since we have several independent assessments, we plan to use a test of the multivariate correlation between the cognitive measures and age of onset. Follow-up tests will be performed with bivariate correlations. Since many of these variables will be non normal, we plan to use standard nonparametric analogs of the multivariate and bivariate correlation. Assuming a conservative significant threshold of 0.01, we will need approximately 110 subjects to achieve reasonable power (i.e., 0.80) to detect moderate to small correlations (r = 0.3). Only 11 more subjects would be needed to achieve the same power using the nonparametric approach. Thus, this project will have adequate power during the second year of assessment. Accurate estimates of the relative value of different measures will need somewhat larger samples and be achieved in years 3 to 4 (with an approximate 95 percent confidence interval ± 0.10).

Hypothesis 2 suggests that the measures of cognitive dysmetria will correlate with frontal lobe function more so than will other measures of focal cortical areas. Partial correlations will be used to correlate the cognitive dysmetria indices with frontal lobe functioning measures while controlling for several variables measuring distal focal cortical areas. These correlations will assess the amount of unique contribution of the cognitive dysmetria variables. Again, we will use either parametric (Pearson's r) or nonparametric (Kendall's tau-b) to assess the partial correlations.

Hypothesis 3 determines how well the cognitive dysmetria measures predict later outcome using the longitudinal data base. Outcome measures will include symptom variables, e.g., (SANS/SAPS), living arrangements, rehospitalization. The basic analysis strategy is appropriate for regression analysis. Simple linear regression, the nonparametric analog, and logistic regression will be used as appropriate. Since Hypothesis 3 also speculates that the cognitive dysmetria measures will provide better predictors than will traditional measures of clinical neuropsychology, tests of correlated correlations will also be performed.

Back to Top

PROJECT 2: COGNITION ASSOCIATED WITH CAUDATE AND CEREBELLAR LESIONS
Project Director: Robert G. Robinson, M.D.
Co-Investigators: Jane S. Paulsen, Ph.D., Jane Springer, Ph.D., Daniel S. O'Leary, Ph.D., Nancy C. Andreasen, M.D., Ph.D.
Funding Status and Proposed Duration: This project is funded by an RO1 grant to Dr. Robinson entitled, "Emotional Regulation of Patients with Brain Injury." This project is funded 1995-2000.
Specific Aims
1. To examine the association among cognitive deficits and the perception and expression of emotion following a mood-inducing video film in normal elderly controls and patients with brain dysfunction.
2. To compare patients with focal caudate lesions, patients with focal cerebellum lesions, patients with focal prefrontal cortex lesions, and patients with psychoses on various clinical and experimental cognitive tasks.
3. To compare performances on tasks of procedural learning with and without feedback in patients with caudate and cerebellar lesions and patients with psychoses.

Background and Rationale

Mood Disorders. There is a body of literature suggesting a right-hemisphere superiority for the comprehension and expression of facial and gestural emotional stimuli, as well as for the comprehension and expression of emotional prosody. Although more controversial, there is also a large literature supporting the hypothesis that there is a hemispheric asymmetry in the regulation of positive and negative emotion. Although the neuroanatomy of emotional regulation is not fully understood, the medial and ventral lateral limbic structures may constitute an integrated system responsible for emotional arousal and motivation, identification of the emotional significance of the stimulus configuration and regulation of the behavioral response. Studies of brain lesions and emotion have found that lesions of the dorsal lateral prefrontal cortex and left anterior basal ganglia have been implicated in the production of depression. The orbitofrontal and dorsolateral frontal cortex have major afferent connections with the caudate, the anterior temporal cortex, the inferior parietal lobe, the cingulate, and the amygdala (Nauta and Domesick 1984). Patients with lesions of the basal ganglia are noted to have a marked loss of activity and motivation. Alexander and DeLong (1986) have identified at least five important basal ganglia thalamocortical loops involving the frontal lobes. Each loop is segregated and parallel to the others and is centered around a separate part of the frontal cortex. These circuits have been designed; motor, oculomotor, dorsolateral, orbitofrontal, and anterior cingulate. With the exception of the motor loop projecting to the putamen, they all project to the caudate, sometimes with projection into the ventral nucleus accumbens. Basal ganglia connections project to the dorsal medial thalamic nucleus and back to the prefrontal cortex. If behavioral and emotional symptoms result from disruption of one or more of these circuits, then disruption of any part of the circuit should lead to the same symptoms. Thus, our selection of caudate and cerebellar lesions is based on these anatomical connections and our desire to study small focal lesions which will not disrupt primary motor pathways.

Cognitive Skills Associated with Caudate and Cerebellar Dysfunction. Since the late 19th century the basal ganglia and cerebellum have been primarily associated with motor control. As much as three decades ago, however, Denny-Brown (1962) described attenuated "stimulus bound" or "environmental dependency" in monkeys with bilateral lesions of the head of the caudate. Since that time, it has been well established that damage to the caudate nuclei resulted in difficulties in the formulation of internal response strategies, relying instead on environmental cues and previously established response tendencies. Although the presence of non-motor frontostriatal circuits is now well recognized, there is little agreement with regard to which functional behaviors are associated with dysfunction of which discrete circuits.

Cohen and Squire (1980) proposed that there were essentially two memory domains, the declarative and the procedural. This view was based on the observations that patients with temporal lobe resections and others with limbic-diencephalic lesions, could learn and perform certain visuomotor or perceptual skills. Mishkin and his colleagues (1984) proposed that the basal ganglia were the most likely anatomical substrate for what they termed the habit system. This more global term encompasses the procedural domain. Butters and his colleagues (Heindel et al 1989; Paulsen et al 1994) were the first group to show a double dissociation, linking skill learning to the basal ganglia in the absence of declarative learning deficits in patients with Huntington's disease, and the converse with regards to traditional amnestic syndromes such as Korsakoff's and Alzheimer's disease. Thus, within this framework it is becoming increasingly clear that even patients with the most "classic" basal ganglia motor disorders, such as Parkinson's and Huntington's disease, suffer from some degree of cognitive impairment.

DeLong and Georgopoulos (1981) suggested that there were two distinct loops or circuits passing through the basal ganglia. The first, a "motor loop," is centered upon the putamen which receives input from sensorimotor cortex channeling output to premotor areas of the frontal cortex. The second association, or "complex loop" receives input from cortical association areas, passes initially through the caudate, finally channeling output back to the prefrontal cortex. In this view, the basal ganglia integrate diverse inputs from the entire cerebral cortex and funnel these influences, via the thalamus, to motor or association circuits of the frontal cortex. Recent developments have begun to clarify and further refine anatomical and behavioral aspects of this model. When viewed as a whole, these basal ganglia-thalamo-cortical circuits appear to play a modulating role in a wide range of behaviors. Broadly speaking, processes such as the generation, maintenance, switching, and blending of motor, mental or emotional sets is involved in the functioning of the basal ganglia. A brief review of the human literature of case studies of discrete caudate lesions suggests that patients with dorsolateral caudate involvement show apathy, aspontaneity and diminished initiative; patients with ventromedial caudate involvement show disinhibition, disorganization and impulsiveness, and patients with larger caudate lesions show affective symptoms with psychotic features (Wang 1991; Bokura and Robinson 1997).

Over the past decade there has been an explosion of research articles suggesting that the cerebellum is associated with several cognitive functions (e.g., Akshoomoff et al 1992; Ivry and Baldo, 1992; Daum et al 1993; Middleton et al 1994). Although several cognitive operations have been emphasized, the most typical cognitive skills addressed involve timing functions (Ivry and Keele et al 1989), learning and memory (Appollonio et al 1993), attention (Courchesne et al 1994), and classical conditioning (Topka et al 1993).

Hypotheses
Although there are several hypotheses being tested central to the funded proposal, in addition, the following hypotheses will be testable through the MH-CRC:
1. The cognitive dysmetria measures of flanker facilitation and inhibition will be sensitive, but not specific, to impairment of the prefrontal cortex, as demonstrated by impairments in all patient groups.
2. The acquisition of prism adaptation and tracking efficiency will be sensitive, but not specific, to impairment of the prefrontal cortex, as demonstrated by impairments in all patient groups. The retention of prism adaptation and tracking efficiency, however, will demonstrate the greatest impairment in patient groups with basal ganglia impairment (e.g., caudate lesions, and schizophrenia patients with tardive dyskinesia).
3. RTC will be most severely impaired in patients with dysfunction of the basal ganglia (caudate lesions and schizophrenia patients with tardive dyskinesia) and cerebellum (schizophrenia and cerebellar lesions) whereas time perception will be most severely impaired in patients with dysfunction of the cerebellum (schizophrenia and cerebellar lesions).

Methods
Subjects: Subjects will be both male and female and between the ages of 50 and 80. Normal controls who are healthy without history of cerebrovascular disease or major psychopathology who are comparable to the patient groups in age, gender, and socioeconomic status will be studied. Subjects with strokes will be obtained from several sources which we have developed over the past two years. Our data collection over the past two years has revealed that we can recruit from a sample of over 500 patients with acute stroke. Thus, we anticipate being able to enroll 11 subjects with caudate lesions and six subjects with cerebellar lesions each year of the five-year study.

Procedures: We will assess in normal elderly subjects or in patients with single lesions of the right or left caudate or cerebellum, the emotional and biological effect of emotional stimulation. After obtaining informed consent, patients will be administered psychiatric, neuropsychological, and neurological exams. The next day, they will undergo an MRI and PET study. Patients will be tested for emotional activation in a quiet room while undergoing PET imaging and having their facial expressions filmed. They will be attached to an EEG and ANS measuring devices and then shown a neutral emotional content film through a video monitor. Patients will then be presented with happy, sad, and fearful films randomly presented, followed by another neutral film. There will be five minutes between sessions when the subjects will report their maximal emotional feeling elicited by the film. Another set of films eliciting the same emotions will be presented to assess for reliability of response, while trying to avoid a habituation decline in emotional response.

Measures: The Present State Examination will be used to evaluate psychiatric symptoms (Robinson et al 1983). The Hamilton Depression Rating Scale (Hamilton 1960), the Zung Self-Rated Depression Scale (Zung 1965), and the Hamilton Rating Scale for Anxiety (Hamilton 1959) will also be administered. An MRI scan will be obtained for each individual in the study at three to four months post stroke. The Johns Hopkins Functioning Inventory (Robinson and Szetela 1981) will be used to assess the patients' ability to perform activities of daily living, such as the ability to dress oneself, read, write, and express needs.

The neuropsychological battery is as follows:
Orientation: Temporal and spatial orientation
Language: Boston Naming Test, Reading, Token Test, Sentence Repetition, Verbal Fluency
Remote Memory: WAIS-R Information, Famous Faces Test
Verbal Memory: Rey Auditory Verbal Learning Test
Visual Memory: Benton Visual Retention Test
Visuoconstruction: WAIS-R Block Design
Executive Functions: Luria Motor Sequences, Verbal and Design Fluency, Wisconsin Card Sorting Test
Premorbid IQ Estimation: WAIS-R Vocabulary
In addition to standardized clinical neuropsychological assessment, we have also developed a battery of examinations to assess symptoms of denial, neglect, apathy, anosognosia, pathological crying, and aprosodia. All subjects will be administered this comprehensive psychiatric functional and cognitive evaluation prior to the PET study.

As described above, the battery to assess cognitive dysmetria will also be administered to this group of patients.

Data Analysis:
All three hypotheses in this project suggest patterns of specificity and sensitivity among the groups of patients and the normals. Thus the analytic strategy will be similar. Basically, a one-way ANOVA (or Kruskal-Wallis test) will be used. One set of tests will contrast the collection of patient groups to the normal control group assessing the sensitivity of the cognitive dysmetria measures. Subsequent tests will also look at specificity in 1) a general sense by reassessing group differences only among the patient groups, and 2) specific measures (e.g., measures of basal ganglia) will be impaired for only certain groups (e.g., schizophrenia patients with TD and patients with caudate lesions) assessing these measures specificity. Since several variables are included within the cognitive dysmetria measures, omnibus tests will be performed for the more general hypotheses with MANOVA.

Back to Top

PROJECT 3: COGNITION AND HUNTINGTON'S DISEASE
Project Director: Jane S. Paulsen, Ph.D.
Co-Investigators: Robert G. Robinson, M.D., Jane Springer, Ph.D., Nancy C. Andreasen, M.D., Ph.D., Daniel S. O'Leary, Ph.D.
Funding Status and Proposed Duration: This project is partially funded by a multi-site R01 grant to Dr. Paulsen entitled "Coenzyme Q10 and Remacemide in Huntington's Disease." This project is funded 1996-2001
Specific Aims
1. To examine the cognitive correlates of early disease in gene positive nonsymptomatic persons at-risk for Huntington's disease and persons with a recent diagnosis of Huntington's disease
2. To measure the cognitive, functional, psychiatric, and neurological correlates of disease progression over 30 months in patients with Huntington's disease assigned to one of four treatment groups: Placebo, Remacemide, CoEnzyme Q10, Remacemide and CoEnzyme Q10.

Background and Rationale
Huntington's disease (HD) is an autosomal dominant neurodegenerative disorder which results from an unstable expansion of the trinucleotide repeat CAG in the IT 15 gene located near the telomere of the short arm of chromosome 4 (The Huntington's Disease Collaborative Research Group 1993). The clinical features of HD usually emerge in adulthood with disorders of voluntary movement and chorea. The disease is relentlessly progressive and affects cognition and behavior as well as motor control, leading to profound functional disability and death over a period of ten to twenty-five years (Greenamyre and Shoulson 1994). The discovery of the HD gene has enabled more widespread testing of asymptomatic individuals at risk for the illness, creating a population of pre-symptomatic gene carriers who await a nearly certain tragic fate. The pathology of HD is characterized by diffuse brain atrophy with severe neuronal loss and gliosis occurring selectively in the caudate nucleus and putamen (striatum). Although there is no treatment for HD at present, several groups are currently pursuing experimental therapeutic and surgical interventions to slow the rate of disease progression or to delay the age of disease onset. As successful treatments become available, greater precision in onset of disease will become essential to establish the ideal time for intervention. It is problematic, however, that the diagnosis of HD is not made until the motor symptoms of the disease are recorded by neurological examination, because several investigators have noted that psychiatric and/or cognitive symptoms may predate the presentation of motor symptoms by several years. If this clinical observation is accurate, initiation of intervention to slow progression or delay onset may be initiated "too late" if onset of motor symptoms is used as the indicator. A more precise measure of disease onset is needed to best determine the appropriate time for intervention.

Neuropsychological Performance in Early HD: Studies of cognition in persons "at risk" for HD have produced variable results. Some studies suggest that persons at risk clearly perform worse on tests of cognition (Diamond et al 1992; Foroud et al 1995; Jason et al 1988; Lyle and Gottesman 1977) whereas other studies show no differences between gene-carriers and non-gene carriers (Strauss & Brandt 1990; Giordani et al 1995). Many studies' findings reside in between these two extremes, however. For instance, Blackmore et al (1995) found that differences between groups were not robust enough to reach traditional statistical significance but did achieve significance using nonparametric analyses. Similarly, Fedio et al (1979) found that mean performances were clearly worse in the gene carriers although findings did not achieve statistical significance. It is important to note that Bradshaw et al (1992) reported that performances were significantly impaired in some persons at-risk for HD (and not others), suggesting that mean comparisons may skew individual differences. There are multiple limitations in the studies conducted to date, however, which confine the conclusions that can be drawn from this body of work. First, there was no way to determine whether the sample consisted of gene-carriers who were close to onset of the disease. (This methodological limitation could explain the heterogeneity of findings reported in the literature thus far.) Second, there was no assurance of gene status (most studies were conducted before the identification of the gene in 1993). Third, the sample sizes were small (often less than 12 per group). Fourth, the cognitive measures varied from study to study and oftentimes normative standards were not considered.

There are two studies which offer methodological advantages over other studies and are worthy of specific mention here. First, Lyle and Gottesman (1977) conducted a retrospective follow-up study of persons at-risk for HD who were tested 15-20 years previously with neuropsychological tests. Results demonstrated that premorbid deficits in intellectual abilities were evident with increasing proximity to the diagnosis of HD, suggesting that the HD gene had been having its effect for years before choreic movements began to appear. In addition, a much larger, and more recent study conducted on 394 at-risk individuals whose gene status was confirmed with DNA data suggested that gene carriers performed worse than non-gene carriers on every test administered, and that mean differences achieved significance on tasks sensitive to speed and sequencing (Foroud et al 1995).

The time is ripe for a large, controlled, comprehensive study of cognitive and behavioral change in presymptomatic gene-carriers for HD. Relative to previous investigations of persons at-risk for HD, methodology can be improved in the following ways: a) gene status can be confirmed by DNA; b) variations in gene repeat size can be used as a modifying influence on age of onset; c) subjects can be selected based upon the probability of imminent HD onset; d) cognitive and behavioral measures can be selected which are specifically sensitive to changes in the caudate and its connections; e) larger sample sizes can be obtained.

Hypotheses
Although there are several hypotheses being tested central to the funded proposal, in addition, the following hypotheses will be testable through the MH-CRC:
1. The cognitive dysmetria measures of flanker facilitation and inhibition will be sensitive, but not specific, to impairment of the prefrontal cortex, as demonstrated by impairments in all patient groups.
2. The acquisition of prism adaptation and tracking efficiency will be sensitive, but not specific, to impairment of the prefrontal cortex, as demonstrated by impairments in all patient groups. The retention of prism adaptation and tracking efficiency, however, will demonstrate the greatest impairment in patient groups with basal ganglia impairment (e.g., Huntington's disease and schizophrenia patients with tardive dyskinesia).
3. RTC will be most severely impaired in patients with dysfunction of the basal ganglia (Huntington's disease and schizophrenia patients with tardive dyskinesia) and cerebellum (schizophrenia) whereas time perception will be most severely impaired in patients with dysfunction of the cerebellum (schizophrenia).

Methods
We propose a double-blind, placebo-controlled investigation of CoQ, remacemide hydrochloride, and combined CoQ/remacemide hydrochloride in 340 ambulatory patients with HD who are enrolled by 22 HSG investigators. Eligible subjects will be randomized to one of four treatment arms using a 2 x 2 factorial design: (1) placebo, (2) CoQ alone, (3) remacemide alone, and (4) the combination of CoQ and remacemide. Comprehensive psychiatric, neurological and cognitive evaluations will occur at 1, 4, 8, 12, 16, 20, 25, 30, and 31 months after a baseline evaluation.

Subjects
Inclusion Criteria:
1. Huntington's disease defined as a characteristic movement disorder in the setting of a confirmatory CAG repeat expansion (> 39) consistent with HD;
2. stages I or II of illness (TFC > 7) wherein patients are ambulatory and do not require skilled nursing care;
3. age of 14 years or older;
4. women who have childbearing potential (i.e., are not postmenopausal or surgically sterile) may participate provided they are using, in the investigator's opinion, adequate birth control methods (e.g. taking highly effective hormonal contraceptives or using an IUD)
5. patients (or legal guardians for minors) must be capable of providing informed consent and complying with the trial procedures;
6. for patients who are taking concurrent psychotropic medications (including antidepressants and neuroleptics), the dosage of these medications should be stable for 4 weeks prior to randomization and should be maintained at constant dosage throughout the 30 months of the study. If clinical conditions mandate modifications of such medications, these changes will be systematically recorded and the subject will be permitted to remain in the trial;
7. patients who have participated in prior studies involving CoQor remacemide will be eligible, although patients must stop taking either medication for at least 3 months prior to enrollment.

Exclusion Criteria:
1. clinical evidence of unstable medical or psychiatric illness;
2. women who are breastfeeding, or have a high likelihood of pregnancy (e.g., inadequate contraception);
3. history of serious alcohol or drug abuse within the previous year;
4. patients with known sensitivity or intolerability to the interventions;
5. patients who have taken any investigational drug within 30 days of baseline assessment and randomization.

Clinical Assessments
Outcome Measures: The primary outcome measure will be the change in total functional capacity (TFC) between the baseline and the 30-month visits. The TFC, a measure of functional disability, is a valid and reliable measure of disease progression. Several secondary clinical efficacy measures will be derived from the UHDRS, including the total motor score, total behavior score, verbal fluency test (Benton and Hamsher 1976), Symbol Digit Modalities Test (Smith 1973), Stroop Interference Test (Stroop 1935), functional checklist score, and Independence Scale score.

In addition to the cognitive measures administered as part of the Unified Huntington's Disease Rating Scale (UHDRS), subjects will be administered a neuropsychological test battery at the baseline and 30-month (or last) visits. The cognitive assessment is an important secondary outcome measure in this clinical trial because cognitive impairment and behavioral dysfunction are cardinal clinical features of HD and the proposed experimental treatments may affect cognition and behavior differently than functional capacity or motor function.

The cognitive test battery was developed by the Neuropsychology subcommittee of the HSG and was designed to meet the following criteria: (1) tests which have been reported to detect specific cognitive dysfunction associated with HD, (2) tests which have greater sensitivity to detecting change over a relatively short period of time, and (3) tests that could be easily and reliably administered by a non-neuropsychologist in a relatively brief amount of time. The cognitive battery is estimated to take approximately 30 minutes to complete and includes:
a. The Hopkins Verbal Learning Test (HVLT) (Brandt 1991), a brief verbal list-learning test with three learning trials, a delayed recall trial, and a recognition trial. Measures of explicit recall, recognition, and response bias are generated. This test is sensitive to learning impairment in HD (Brandt et al 1992).
b. The Brief Test of Attention (BTA) (Schretlen et al in press) is a brief audio cassette tape-based test of divided verbal attention that has been shown to be sensitive to attentional deficits in HD patients (Schretlen et al in press). In the OPC-14,117 trial, this test revealed a trend toward a beneficial effect of treatment.
c. The Trail Making Test (TMT) (Reitan and Wolfson 1958) is a test of visual-motor speed and of simple and alternate sequencing that has been well documented to demonstrate impairment in HD patients (Starkstein et al 1988; 1992).
d. The Conditional Associative Learning Test (CALT) (Petrides 1990) is a nonverbal test of visual associative learning. Subjects learn to pair each of four spatially disparate unmarked stimuli with one of four unmarked response cards over multiple trials using a trial and error approach. Total number of errors and the number of trials required to complete 12 consecutive errorless trials are obtained. In the OPC-14,117 trial, this test showed the strongest trend toward a beneficial treatment effect.
In addition to the cognitive test battery, we plan to obtain behavioral information on subjects from their designated caregiver. The primary behavioral measures of interest focus on affective changes and personality changes known to be associated with the frontostriatal dysfunction which occurs in HD. These include:
e. Hamilton Depression Inventory (HDS) Informant Interview. The HDS is a measure of depressive symptomatology. This scale was developed for use in populations in which the subject may not be a reliable informant of psychiatric changes (Hamilton 1960).
f. Frontal Lobe Personality Scale (FLOPS). The FLOPS is a self-report questionnaire completed by the patient's primary caregiver. The FLOPS assesses the cardinal features of behavioral disturbance reported in HD, including disinhibition/distractibility, apathy, and self-monitoring (Cummings 1995; Grace and Malloy 1992).
As described above, the battery of cognitive dysmetria will also be administered to this group of patients.

Data Analysis
The analyses here are essentially the same as for Project 3. Briefly, a series of ANOVA (or Kruskal-Wallis tests) will be used to assess group differences. Overall tests will be performed with MANOVA to help control Type I errors. We have specific hypotheses regarding several variables (e.g., retention of motor memory) which we expect to show specific deficits in the Huntington group. Specific contrasts in the context of the ANOVA's will be used as direct tests for the hypotheses. For this analysis, we will also be sensitive to the stage of illness in the patient group using dementia severity as a covariate.

Back to Top

PROJECT 4: COGNITION, PSYCHOSIS AND ALZHEIMER'S DISEASE
Project Director: Jane S. Paulsen, Ph.D.
Co-Investigators: Robert G. Robinson, M.D., Jane Springer, Ph.D., Nancy C. Andreasen, M.D., Ph.D., Daniel S. O'Leary, Ph.D.,
Funding Status and Proposed Duration: This project is funded by an R29 grant to Dr. Paulsen, entitled "The Neuropsychology of Psychosis in Alzheimer's Disease." This project is funded 1996-2001
Specific Aims
1. To determine the cross-sectional and longitudinal differences between matched groups of psychotic (ADpsy+) and nonpsychotic (ADpsy-) patients with a diagnosis of AD on a number of neuropsychological measures. Recent evidence suggests that severe limbic and frontostriatal dysfunction may mediate psychotic symptoms in AD. If this is the case, AD patients with psychosis may demonstrate particularly severe deficits on neuropsychological tests of frontostriatal dysfunction relative to equally globally demented AD patients without psychosis. If the presence of psychotic symptoms is related to unusually severe limbic and frontostriatal dysfunction in AD patients, then psychotic symptoms may also be predictive of greater functional and cognitive decline.
2. To examine the cross-sectional levels of functional impairment and longitudinal rates of functional decline across ADpsy+ and ADpsy- groups. Psychotic symptoms may interfere with the performance of such activities of daily living as grooming, shopping and appropriate social interactions, despite sufficiently preserved cognitive capacity to perform these functions. In addition, evidence suggests that AD patients with delusions have greater functional decline. ADpsy+ patients may require earlier institutionalization than ADpsy- patients due to greater functional impairment, or because behavioral manifestations of their psychosis make them unmanageable at an earlier stage of the disease.
3. To identify associated clinical correlates of psychosis in AD, such as extrapyramidal signs (EPS), family history of psychiatric disorder, and/or clinical diagnosis of Lewy Body Variant (LBV). We anticipate that ADpsy+ will demonstrate a higher incidence of EPS, family psychiatric history, and clinical diagnosis of LBV than ADpsy- patients.
4. To identify associated neuroanatomical correlates of ADpsy+, as demonstrated by greater pathology on MRI and postmortem measures of frontal and striatal brain regions in ADpsy+ than in ADpsy. We anticipate that ADpsy+ will have less frontal and basal ganglia volume on quantitative MRI than ADpsy-. In addition, we anticipate that ADpsy+ will have a greater number of Lewy bodies on neuropathological evaluation.
5. To evaluate response to pharmacologic treatment of AD-associated symptoms. We anticipate that (a) ADpsy+ patients will demonstrate greater susceptibility to tardive dyskinesia secondary to neuroleptic treatment (due to greater dopamine and less acetylcholine) than ADpsy- patients; and (b) ADpsy+ patients will demonstrate fewer beneficial effects of Cognex than ADpsy- patients.

Background and Significance
The identification, diagnosis, and management of psychotic symptoms in AD are of preeminent importance in helping practitioners manage AD patients, and facilitating the quality of life in AD patients and their caregivers. However, few long-term prospective investigations have examined the risk factors, phenotypic characteristics, potential underlying mechanisms, and course of psychosis in AD. The current proposal improves upon previous studies by combining sound methodology (i.e., case definition, subject source, subject selection, sample size, and selection of controls), standardized psychiatric rating scales, thorough standardized neuropsychological assessment, and data analyses which considers confounding factors in a prospective research design. Although an excellent study was conducted by Stern et al (1994) there were no standardized neuropsychological or psychiatric assessments obtained to evaluate the hypotheses stated in the present proposal. Nearly all previous investigations of cognition in AD patients with psychosis have relied upon a single global score (e.g., MMSE) which precludes the possibility of identifying discrete neuropsychological differences between these groups. In summary, the available literature has shown the following: 1) that psychosis in AD is associated with a greater frequency of behavioral problems, aggression, and hostility; 2) that psychosis in AD is associated with earlier institutionalization; 3) that psychosis in AD is one of the most serious problems affecting quality of life for AD and caregivers; 4) that ADpsy+ patients show a more rapid rate of decline as measured by the MMSE; 5) that the presence of delusions is associated with greater functional impairment; 6) that AD patients with Lewy Bodies and/or EPS have a greater incidence of psychosis; and 7) that "frontal lobe" performance is worse in AD patients with psychosis. Much remains to be done to combat this important public health problem. First, no previous study has used standardized psychiatric rating scales to characterize the types of psychiatric symptoms in AD and to explore large patient groups for subtypes. Second, no study has included comprehensive neuropsychological evaluation in the investigation of psychosis in AD. Third, only the most recent studies have begun to consider the Lewy Body hypothesis of psychosis in AD and few studies have combined these areas of investigation. Fourth, most of the findings cited have not been replicated with larger subject samples. The proposed studies will advance our knowledge of the frequency, neuropsychiatric associations, course of, risk factors for, and potential underlying mechanisms associated with psychosis in AD. The findings are also likely to have implications for the treatment of AD patients with psychosis, as well as for understand-ing the neuropathology of psychotic symptoms in AD.

Hypotheses
Although there are several hypotheses being tested central to the funded proposal, in addition, the following hypotheses will be testable through the MH-CRC:
1. The cognitive dysmetria measures of flanker facilitation and inhibition will be sensitive, but not specific, to impairment of the prefrontal cortex, as demonstrated by impairments in all patient groups.
2. The acquisition of prism adaptation and tracking efficiency will be sensitive, but not specific, to impairment of the prefrontal cortex, as demonstrated by impairments in all patient groups. The retention of prism adaptation and tracking efficiency, however, will demonstrate the greatest impairment in patient groups with basal ganglia impairment (e.g., 3. Alzheimer's disease with psychoses and schizophrenia patients with tardive dyskinesia).
3. RTC will be most severely impaired in patients with dysfunction of the basal ganglia (psychotic Alzheimer's disease and schizophrenia patients with tardive dyskinesia) and cerebellum (schizophrenia) whereas time perception will be most severely impaired in patients with dysfunction of the cerebellum (schizophrenia).
Methods
Subjects We plan to recruit 100 AD patients with current psychosis and 100 AD patients without past or current psychosis during the first 3 1/2 years of the proposed study. We expect a drop-out rate of approximately 15% per year. Subjects who drop-out and those who develop psychosis de novo during the course of the study will be replaced for statistical analyses of group comparisons (e.g., psychotic AD versus nonpsychotic AD) although their data will be used for analyses of other hypotheses.
Inclusion Criteria
a. Fluent in speaking English (whether or not it was the first language) prior to the onset of dementia.
b. Diagnosis - We will use DSM-IV (APA 1994) and NINCDS-ADRDA (McKhann et al 1984) criteria for AD, DSM-IV for any other psychiatric diagnosis, and ICD-9 categories (World Health Organization 1978) for other medical diagnoses.
c. Physically and psychiatrically stable enough to undergo the various assessments.
d. Presence of medical records and a "significant other" to corroborate history and ensure follow-up.
Exclusion Criteria
a. Clinical evidence of focal neurologic disorders.
b. History of head injury with loss of consciousness > 30 min.
c. History of DSM-IV alcohol dependence or illicit drug use within the last two years.
d. Other axis I disorders (DSM-IV criteria) at present.
Baseline Evaluation
(A) Demographic, Medical, Pharmacologic and Physical Exam Data:
A clinical diagnosis of possible Lewy Body Variant of AD is given to a patient following a neurological examination during which the neurologists observes two out of the following four clinical signs of parkinsonism: masked facies, stooped posture, shuffling gait, and/or tremor. We will keep a complete record of all the psychotropic and nonpsychotropic medications (name, daily dose, duration, indications, therapeutic and adverse effects, and compliance) for every subject at every visit.
(B) Neuropsychiatric Assessment
1. Psychiatric Ratings: The Neurobehavioral Rating Scale, or NRS (Levin et al 1987), is a 27 item multidimensional tool for the assessment of psychopathology, which includes most of the items of the Brief Psychiatric Rating Scale or BPRS (Overall and Gorham 1962). The Neuropsychiatric Inventory, or NPI (Cummings et al 1994) is a relatively new instrument designed specifically to assess 10 behavioral disturbances occurring in dementia patients: delusions, hallucinations, dysphoria, anxiety, agitation, euphoria, disinhibition, irritability, apathy, and motor activity.
2. Functional Ratings: Activities of Daily Living will be assessed using four measures of family-rated, patient-rated, and staff-rated independence. The Pfeffer Outpatient Disability Scale (Pfeffer et al 1982) is a 10-item family-rated scale which evaluates the patient's level of independence in activities of daily living. The Lawton-Powell Physical Self-Maintenance Scale (Lawton et al 1969) is a 6-item scale of independence in activities of daily living as perceived by the patient. The Nurses Observation Scale for Geriatric Patients is a 35-item scale of ADLs as judged by professional staff. The AD Dependency Scale is a 13-item scale which has been used in the AD Cooperative Studies of Clinical Trials in AD and has been shown to be sensitive to subtle changes in functional ability.
3. Movement Ratings: The modified Simpson-Angus Neurological Rating Scale (Simpson and Angus 1970) is an 8-item scale which assesses parkinsonian symptoms such as tremor, rigidity and abnormal gait. The modified Abnormal Involuntary Movement Scale (AIMS; Guy 1976) is an 18-item scale which assesses involuntary movements.
1. Cognitive Assessment
Dementia Severity: 1) Mini-Mental State Examination(MMSE) (Folstein et al 1975); 2) Mattis Dementia Rating Scale(DRS) (Mattis 1976)
Expressive Language: 1) Category and Letter Fluency; 2) Boston Naming Test (BNT); 3) Writing.
Receptive Language: 1) Auditory Comprehension; 2) Reading; 3) Token Test - Part VI.
Attention: 1) Digit Span; 2) Visual Span; 3) Continuous Performance Test (CPT)
Learning: 1) California Verbal Learning Test (CVLT) Trials 1-5; 2) Story Learning; 3) Figure Learning
Delayed Recall: 1) CVLT: Short Delay vs. Trial 5; 2) Story Memory; 3) Figure Memory
Abstraction/Executive Functions: 1) Trails B; 2) Wisconsin Card Sorting Test 3) WAIS-R Similarities
Visuospatial: 1) Clock Drawing (CDT) and Clock Setting; 2) Crosses; 3) Block Design
Psychomotor Speed: 1) Trails A; 2) Digit Symbol
Motor Ability: 1) Grooved Pegboard; 2) Finger Tapping; 3) Luria 3-step
Procedural Learning: 1) Prism Adaptation (PA); 2) Rotor Pursuit (RP); 3) Weather Test (WT)
As described above, the battery of cognitive dysmetria will also be administered to this group of patients.

Data Analysis
The hypotheses in this project involve 5 groups: Normal controls (NC), schizophrenic patients with TD (STD) and those without TD (SNTD), and Alzheimer's patients with psychosis (AP) and without (ANP) psychosis. Each of the 3 hypotheses in this project suggest different patterns of sensitivity and specificity among the groups on the cognitive measures. For instance, we speculate that retention of motor memory will be impaired only with STD and AP groups. Simple ANOVA (or Kruskal-Wallis) tests followed by specific 1 df contrasts will be used to directly address these hypotheses. Either simple -tests or Mann-Whitney tests will be used. For these specific single contrast hypotheses we will have sufficient power (power = 0.80) to detect effect sizes as small as Cohen's d = 0.33 (significance threshold = 0.05) by the end of the second year of study.

Back to Top

PROJECT 5: COGNITION AND MOOD DISORDERS IN BRAIN INJURY
Project Director: Robert G. Robinson, M.D.
Co-Investigators: Jane Springer, Ph.D., Jane S. Paulsen, Ph.D., Daniel S. O'Leary, Ph.D., Nancy C. Andreasen, M.D., Ph.D.
Funding Status and Proposed Duration: This project is funded by an R01 grant to Dr. Robinson, entitled "Mood Disorders following Traumatic Brain Injury." This project is funded 1995-2000.
Specific Aims
The overall aim of this project is to characterize mood and cognition following traumatic brain injury (TBI) More specific aims are:
1) To determine the cognitive correlates of mood disorders following traumatic brain injury (TBI);
2) To examine the association among cognition, mood and functioning as determined by activities of daily living, occupational and psychosocial ratings following TBI;
3) To determine the effect of pharmacological treatment of depression on mood, cognitive recovery, and functioning.

Background and Rationale
The annual incidence of traumatic brain injury in the U.S. has been estimated to be approximately 500,000 persons per year. Approximately one-fifth (96,000) of these patients die, and a similar number face severe and chronic disability at a cost of $12.5 billion annually (Frankowski 1986). TBI does not appear to be decreasing in incidence despite greater preventive efforts. In addition, progress in acute medical and surgical care are increasing the number of TBI survivors. These patients are often left with significant impairment, including not only sensory and motor deficits but also higher-level cognitive, affective and social dysfunction. How type and severity of brain injury relate to the ultimate functional outcome, along with factors that influence the extent and rate of recovery, all require further research.

The long-term outcome of TBI patients is primarily related to severity of brain injury. In addition, the type and location of the intracranial lesion, as well as the efficacy of acute medical and surgical treatments, may have a decisive impact on recovery (Bullock and Teasdale 1990; Gennarelli et al 1982; Levin et al 1990). Outcome is also influenced by concurrent factors that include age (Vollmer et al 1991), socioeconomic status, educational level, previous psychiatric disorders (e.g., history of alcohol and/or drug abuse, personality disorders, etc.), and premorbid levels of social functioning (Lishman 1973). Finally, the quality and extent of rehabilitation services and the availability of social and vocational support also play a significant role in TBI outcome.

Few studies have examined the prevalence of mood disorders (e.g., major depression or mania) in TBI patients and their effect on outcome variables (Robinson and Jorge 1994). Dikmen and Reitan (1977) reported on the longitudinal evolution of emotional functions (measured using MMPI scores) in a consecutive series of 27 TBI patients followed for 18 months. They concluded that emotional disturbances tend to decline over time after injury and that they are associated with the presence of greater neuropsychological impairment. On the other hand, Fordyce et al (1983) found significantly higher MMPI depression scores in the chronic stage of TBI (i.e., after 6 months from brain injury) than in the more acute stages. The reported frequency of depressive disorders following TBI has varied from 6 to 77% (Rutherford 1977; Varney et al 1987). McKinlay et al (1981) reported indirect evidence of depressed mood in about half of their patients at 3, 6, or 12 months following severe brain injury. Schoenhuber and Gentilini (1988) found depressive symptoms in 39% of 103 patients with mild head injury interviewed at one year follow-up, and concluded that these patients have an increased risk of developing depression compared with an appropriate control group. Kinsella et al (1988) reported that 33% of 39 patients with severe brain injury were classified as depressed within two years, and 26% as suffering from anxiety. Almost 70% of a convenient sample of 60 married, brain-injured subjects demonstrated at least mild anxiety (Linn et al 1994). Overall, Gualtieri and Cox (1991) estimated that the frequency of major depression in traumatic brain injured patients lies between 25 and 50%.

In summary, although patients with moderate head injuries may present with relatively mild physical impairments, they may experience behavioral disorders that have a significant impact on the extent and quality of their interpersonal relationships, and affect their re-entry into the community. Major depression appeared to have a deleterious effect on both psychosocial and activities of daily living outcomes. Since depressive disorders tend to be resolved by one year, we presume that depression may negatively influence patients' participation in rehabilitation efforts early during their course of recovery, and that they do not recover these early losses even when the depression is over.

Hypotheses
Although there are several hypotheses being tested central to the funded proposal, in addition, the following hypotheses will be testable through the MH-CRC:

1) The cognitive dysmetria measures of flanker facilitation and inhibition will not be significantly different in TBI patients with and without mood disorders. TBI patients with mood disorders will, however, demonstrate a slower average reaction time than TBI patients without mood disorder.
2) The acquisition of prism adaptation and tracking efficiency will be equally impaired in both TBI groups.
3) RTC will be poorer in TBI patients with mood disorders. There will be no difference between TBI groups in Time Perception.

Methods
Patients will be characterized at intake to establish their premorbid level of physical, cognitive, and psychosocial functioning, and to identify premorbid risk factors for psychiatric disorders. The head injury will be classified with regard to lesion morphology and impaired cerebral function, and non-CNS injuries will be categorized.

Inclusion criteria:
Patient with a first closed-head injury diagnosis and classified as having moderate or severe head injury.
Age 18 to 65.
A systolic blood pressure of at least 80mmHg at the moment of the GCS score determination.

Exclusion criteria:
Patients with a penetrating head injury.
Association of traumatic brain injury and significant spinal cord injury that produces to [sic] quadriplegia or paraplegia
Presence of significant physical impairment unrelated to CNS injury occurring at the time of the TBI (e.g., multiple long bone fractures, craniofacial deformities, severe chest or abdominal trauma) that may significantly influence the long-term of [sic] recovery from the traumatic episode. This is intended to avoid patients whose outcome will be markedly influenced by non-CNS injury.
Presence of moderate to severe comprehension deficits related to previous brain damage. This includes patients with a history of alcohol abuse and signs of organic CNS damage, either radiological (e.g., brain atrophy in CT scans) or clinical (e.g., alcoholic polyneuropathy).
Patients with a severe complicating illness such as uncompensated congestive heart failure.

Patients will be evaluated during their hospital admission and at 3, 6, 12, and 24 months after trauma. Neuroimaging will be performed to classify the head injury, and a battery of test instruments will be administered at set intervals to assess physical, cognitive, and psychosocial functioning. Over the next two years, patients will be evaluated at set intervals to characterize their psychiatric status, to monitor rehabilitation services, and to assess functional recovery. Variables to be assessed include the following:
Severity of brain injury.
Type and location of brain damage.
Type and severity of non-CNS related injury.
Quality of social functioning and numbers of social ties.
Personal history of alcohol and/or drug abuse.
Personal history of psychiatric disorders.
Family history of psychiatric disorders.
Degree of physical and cognitive impairment.
Type and severity of neurological impairment, including assessment of neglect, denial, etc.
Type and extent of rehabilitation services

Cognitive function will be measured with the Mini-Mental State Examination (MMSE) (Folstein et al 1975), a 30-point instrument that screens orientation, attention, recall and language functioning. The MMSE has been shown to be a reliable and valid means of assessing a limited range of cognitive functions in several medically-ill or brain-injured populations (Robinson and Szetela 1981). MMSE scores below 24 are indicative of clinically significant cognitive impairment.

Severity of depression and anxiety will be determined using the 17-item Hamilton Rating Scale for Depression (HAMD) (Hamilton 1960) and the Hamilton Rating Scale for Anxiety (HAMA) (Hamilton 1959), respectively. These instruments have proved to be valid and reliable, not only among psychiatric patients, but also among brain-damaged patients (Hamilton 1959). Severity of manic syndromes will be quantified using Young's Mania Rating Scale (MRS) (Young et al 1983). In addition, we will assess the existence and severity of emotional lability using the Pathological Laughing and Crying Scale (PLACS) (Robinson et al 1993). And emotional outburst using the Catastrophic Reaction Scale (CRS) (Starkstein 1993).

Family history information will be obtained from the patient and close relatives using the Family History Research Diagnostic Criteria (FHRDC) (Andreasen 1977). Personal history of alcohol or other substance abuse will be quantified using the Brief Michigan Alcoholism Screening Test (Brief MAST) and the Drug Abuse Screening Test (DAST), whose validity and reliability have been adequately established (Selzer 1971; Skinner 1982). A score of 5 or greater on the MAST has a 98% sensitivity and 95% specificity for the detection of drug abuse. Information regarding personal and family history of alcohol or other substance use will also be requested from close relatives.

Neuropsychological evaluation
A full neuropsychological battery will be carried out initially and at 12 and 24 months following TBI. This neuropsychology battery was developed to assess patients with brain injury, and its reliability and validity have been established (Bolla-Wilson et al 1989). The battery is composed of subsections of the Wechsler Adult Intelligence Test (WAIS) and other standard examinations, and assesses the following range of neuropsychological domains:
-orientation (temporal and spatial);
-language (Boston Naming Test, reading, Token Test, repetition of phrase from Boston Aphasia Battery, and verbal fluency);
-remote memory (general information and Famous Faces Test);
-verbal learning/memory (Rey Auditory Verbal Learning Test);
-visual learning/memory (Benton Visual Retention Test);
-visuoperceptual/visuoconstructional ability (Block Design from the WAIS);
-calculation (WAIS arithmetic subtest);
-executive/motor (Luria Motor Sequences);
-frontal lobe function (verbal and design fluency, Wisconsin Card Sorting Test);
-premorbid I.Q. (estimated based on vocabulary from WAIS).

Additional tests will be administered to assess neglect, anosognosia (denial of illness) and aprosody (lack of response to emotional stimuli), and thus to characterize the scope and specificity of deficits in awareness. In particular, these measures will be used to determine whether damage to specific brain regions differentially affects the perception of emotion.

Assessment of neglect. Our battery of tests to assess neglect includes: 1) visual, auditory, and somesthetic performance after double-simultaneous stimulation; 2) motor neglect; 3) motor impersistence; 4) personal neglect; and 5) hemispatial neglect.

Anosognosia scale. In previous work, we have developed a scale to determine the presence and severity of anosognosia, denial of illness, and established its validity and reliability (Starkstein et al 1993).

Aprosody tests. This battery is designed to evaluate recognition of facial emotion and comprehension of emotional prosody, respectively. Visual perception of emotion will be tested using emotional faces from the Eckman series, and used to determine the presence of visual sensory aprosodia. Verbal sensory aprosodia will be determined by testing the patient's ability to recognize emotional intonation in voices using audio tapes developed by Kenneth Heilman and colleagues. We have used these tapes previously to study aprosodia following stroke (Starkstein et al 1994).

Assessment of Functioning
The functional assessment of TBI patients comprises the following assessments:

-Glasgow Outcome Scale (GOS) (Jennett and Bond 1975). In assessing brain-injured patients, the GOS takes into account survival, level of care needed in daily living, and social re-integration. The 5-category GOS has been widely used since 1975, and has acquired international acceptance.
-Rancho Los Amigos Hospital Levels of Cognitive Functioning (RLCF) (Hagen et al 1979). The RLCF is an 8-category classification of recovery from TBI. It is based primarily on cognitive and behavioral performance with the primary emphasis being memory and agitation.
-John Hopkins Functioning Inventory (JHFI) (Robinson and Szetela 1981). The scale measures functional independence and communicative functions. Scores range from 0 to 27, with higher scores indicating more severe functional impairment. This is a valid and reliable scale for activities of daily living, and has been used extensively in brain-damaged patients.
-Functional Independence Measure (FIM) (Forer et al 1987).