Neurophysiological Targets for Cognitive Training in Schizophrenia

The purpose of this study is to determine whether computer-based training of auditory and visual processing results in corresponding improvement in brain function in individuals with schizophrenia.

Schizophrenia is recognized as one of the leading causes of medical disability worldwide, ranked 9th overall by the World Health Organization, and affects more than 2 million Americans per year. There is considerable evidence to suggest that disability status in schizophrenia relates more directly to cognitive Impairment, involving attention, reasoning, and memory, than to characteristic symptoms of psychosis. Accordingly, the evaluation and advancement of interventions designed to restore cognitive function, generally termed cognitive remediation, is of critical importance to our rehabilitation mission. Recent randomized controlled trials of cognitive remediation in schizophrenia have found moderate gains in cognitive function and improved outcomes in important areas of community living. However, despite these encouraging findings, there remains sparse evidence in support of assumptions that (1) cognitive outcomes represent benefits of training-induced adaptive learning, (2) that training effects are specific to method of intervention, or (3) that change in cognitive test performance occurs through restoration of impaired neural circuitry in schizophrenia. This project will begin to address these issues by examining modality-specific effects of computer-based cognitive training on psychophysiological measures of sensory information processing. Training will be administered using two commercially available computer-based software packages, separately targeting auditory and visually-mediated processes using principles of bottom-up perceptual learning. Two psychophysiological paradigms, mismatch negativity (MMN) and P300 generation, will be administered as tests of early visual and auditory processing. MMN and P300 have been studied extensively in human neuroscience as probes of sensory echoic memory and attention engagement to contextually relevant information. Furthermore, reductions in MMN and P300 generation are reliably observed in schizophrenia, follow the course of a progressive neuropathological process, and correlate with severity of cognitive impairment. The specific aims of this study are to determine: (1) whether training selectively influences bottom-up (MMN) or top-down (P300) information processing, (2) whether training effects are modality (auditory vs. visual) specific, (3) whether baseline MMN and P300 predict, or rate-limit, training progress, and (4) whether pre-post change in cognitive test performance is mediated by neural-level change in MMN and P300 generation. Answers to these questions will provide information needed to structure cognitive training for maximum benefit in schizophrenia.

4 weeks (20 sessions) of auditory cognitive training (Brain Fitness) followed by 4 weeks (20 sessions) of visual cognitive training (Insight)

4 weeks (20 sessions) of visual cognitive training (Insight) followed by 4 weeks (20 sessions) of auditory cognitive training (Brain Fitness)

The program entails six computer-based exercises that are designed to be very easy to use and require no computer experience. The exercises are calibrated to individual performance at the onset of training and adapt in difficulty to individual performance, giving constant feedback about progress. Each of the six exercises focuses on a distinct process: (1) auditory processing speed, (2) discriminating sounds, (3) sound precision, (4) sound sequencing, (5) working memory, and (6) narrative memory. Training was administered in a supervised clinical laboratory setting at a frequency of five 60-minute sessions per week over 4 weeks.

The program entails five computer-based exercises that are designed to be very easy to use and require no computer experience. The exercises are calibrated to individual performance at the onset of training and, following our laboratory procedures, calibration testing is repeated every 5th session. Exercises adapt in difficulty to individual performance, giving constant feedback about progress. Each of the five exercises focuses on a distinct process: (1) visual precision, (2) visual processing speed, (3) divided attention, (4) visual working memory, and (5) useful field of view. Training was administered in a supervised clinical laboratory setting at a frequency of five 40-minute sessions per week over 4 weeks.

Visual P300 was measured in a 3-stimulus target detection task with target stimuli (10%; large circle) presented in pseudo random order amidst a series of novel (10%; fractal), and standard (80%; small circle) images on a 24" LCD monitor at 100cm viewing distance. Subjects are instructed to press a reaction time button with the preferred hand to Targets only, giving equal importance to speed and accuracy. Primary analysis are based on Target "P300b" identified as the most positive amplitude deflection within the window of 250-550ms post stimulus at posterior midline electrode Pz. The P300b component is thought to reflect cognitive processes involved in memory updating and decision making. P300 reported as difference scores from baseline with negative values indicating increased P300.

Auditory MMN is a fronto-central, mid-latency, potential generated by the auditory cortex in response to deviation in a repetitive stimulus sequence. MMN was assessed using a 3-deviant paradigm in which a series of standard tones (633 Hz, 50ms duration,90%) is interrupted by deviants (10%) that differ either by (1) pitch (1000Hz, 50ms), (2) duration (633 Hz, 100ms), or (3) both (1000Hz, 100ms). MMN was tested concurrently with Visual P300 using a combined task in which subjects were instructed to ignore the auditory stimuli and focus on the visual stimuli. MMN is scored by subtracting each deviant ERP waveform from the standard waveform and measuring the most negative deflection in a window of 50 to 265ms post-stimulus from the resulting difference wave. Primary analysis is based on the combined deviant condition scored at the frontal midline (Fz) electrode site. MMN reported as difference scores from baseline with positive values indicating increased MMN.

The Cognitive Composite score is derived from the MATRICS Consensus Cognitive Battery (MCCB). The MCCB consists of 10 tests and provides standard scores for each according to seven cognitive domains: (1) speed of processing, (2) attention/vigilance, (3) working memory (verbal and visual), (4) verbal learning, (5) visual learning, (6) reasoning and problem solving, and (7) social cognition. The primary dependent measures derived from the MCCB for purpose of this study is the cognitive composite score, computed as the average of standard (t-scores) scores from each domain excluding social cognition. MCCB Composite reported as difference scores from baseline with negative values indicating higher test performance.

The visual learning domain of the MCCB is assessed using the Brief Visuospatial Memory Test-Revised (BVMT-R). In three Learning Trials, the respondent views a stimulus display for 10 seconds and is then asked to draw as many of the figures as possible in their correct location on a page in the response booklet. Scores represent overall accuracy across the three trials, with higher scores indicating better learning. Analysis is based on age- and gender-corrected t-scores. BVMT-R reported as difference scores from baseline with negative values indicating higher test performance.

The verbal learning domain of the MCCB is assessed using the Hopkins Verbal Learning Test-Revised (HVLT-R). In three Learning Trials, the respondent listens to a 12-item word list read by an examiner and is then asked to recall as many of the words as possible from memory. Scores represent overall accuracy across the three trials, with higher scores indicating better learning. Analysis is based on age- and gender-corrected t-scores. HVLT-R reported as difference scores from baseline with negative values indicating higher test performance.

Visual working memory is assessed in the MCCB using the Spatial Span task of the Wechsler Memory Scales-III. Using a board on which 10 cubes are irregularly spaced, the examiner taps patterns of increasing length. The respondent is asked to follow by tapping the pattern in the same or reverse sequence. Scores represent total accuracy combined over forward and reverse span conditions. Analysis is based on age- and gender-corrected t-scores. Spatial Span reported as difference scores from baseline with negative values indicating higher test performance.

Auditory working memory is assessed in the MCCB using the Letter-Number Sequencing (LNS) task. LNS is an orally administered test in which and examiner reads strings of numbers and letters, of increasing length over trials, and the respondent mentally reorders the string and reports back to the examiner verbally. Scores represent total number of accurate trials. Analysis is based on age- and gender-corrected t-scores. LNS reported as difference scores from baseline with negative values indicating higher test performance.

Inclusion Criteria: DSM-IV diagnosis of schizophrenia or schizoaffective disorder Age between 18 & 70 minimum of 30 days since discharge from last hospitalization minimum of 30 days since last change in psychiatric medications receiving mental health services no housing changes in the past 30 days Exclusion Criteria: current diagnosis of alcohol or substance abuse history of brain trauma or neurological disease chart diagnosis of mental retardation or premorbid intelligence < 70 based on Wechsler Test of Adult Reading (WTAR) full-scale estimated IQ auditory or visual impairment that would interfere with study procedures a sample of 20 healthy community volunteers was also recruited according to these criteria and tested, without intervention, as a normative reference sample for MMN and P300 measures