RECENT PRESENTATIONS

Gevins, A., Smith, M.E., McEvoy, L.K., Ilan, A., Lazzara, M. (2004). Development of a Clinical Neurophysiological Test of Working Memory. Neurology, 62, Suppl. 5, A529

ABSTRACT

Objective: Develop an efficient neurophysiological test of working memory for assessing treatment-related changes in neurocognitive function.
Background: Neuropsychological tests allow inferences about cognitive ability but do not provide direct information about brain function, or about why a patient's test has performance improved or deteriorated. For instance, deteriorated performance could result from impaired neural processing, diminished alertness and/or simply a lack of effort. Accordingly, we are developing a test of neurocognitive function that combines psychometric tests with simultaneous measures of brain function. Working memory (WM), the conscious process of maintaining and manipulating representations in one's mind for several seconds, is a fundamental cognitive process and was chosen for our initial test. EEGs and EPs, which are sensitive indicators of WM and are more practical than fMRI when 3-D localization is not needed, have been chosen to measure brain function.
Design/Methods: In a series of studies using repeated measures designs, an 8-minute, spatial n-back WM task was performed while EEG was recorded. Using multivariate statistical analysis, task performance and well-studied EEG and EP measures including alpha and theta power and P300 amplitude were compared between a patient's baseline and post-treatment conditions to assess the degree to which neurocognitive function changed.
Results: Prerequisite normative studies have been completed. The WM task-related EEG and EP signals were stable after practice, had high test-retest reliability (N=20; r>.90; p < .001), and were consistent from age 8 to 80 (N=140). The signals had high face validity in that their modulation by variations in WM task difficulty was accurately discriminable in individual subjects (N=8; p < .001). They had high construct validity in that WM task performance and EEG/EP signals were good predictors of individual differences in cognitive ability as measured by the WAIS-R IQ score (N=80, multiple R=.80; p < .001). Finally, they had high discriminative validity in that the WM performance and EEG/EP signals were highly sensitive to transient and chronic changes in neurocognitive function due to a variety of drugs including antihistamines (N=36), caffeine (N=16), alcohol (N=36), marijuana (N=10), anti-epileptic drugs (N=36) and anxiolytics (N=10), as well as to extended wakefulness (N=25) and circadian factors (N=8). Large individual differences were evident in these effects. The combination of task performance and EEG measures was more sensitive to these effects than task performance measures by themselves, and the source of impaired performance was inferable from the EEG measures.
Conclusion: These results suggest that a clinical neurophysiological test of working memory is scientifically feasible. Studies are underway to determine its utility as an outcome measure in clinical studies including treatment of obstructive sleep apnea, aging-related cognitive decline, ADHD and seizure disorders.
Supported by grants from NIMH, NINDS, NIA, NICHHD, NIAAA, NIDA and NHLBI.