Combined Psychophysiological Assessment of ADHD: A Pilot Study of Bayesian Probability Approach Illustrated by Appraisal of ADHD in Female College Students

Abstract

Manifestations of ADHD are observed at both psychological and physiological levels and assessed via various psychometric, EEG, and imaging tests. However, no test is 100% accurate in its assessment of ADHD. This study introduces a stochastic assessment combining psychometric tests with previously reported (Consistency Index) and newly developed (Alpha Blockade Index) EEG-based physiological markers of ADHD. The assessment utilizes classical Bayesian inference to refine after each step the probability of ADHD of each individual. In a pilot study involving six college females with ADHD and six matched controls, the assessment achieved correct classification for all ADHD and non-ADHD participants. In comparison, the classification of ADHD versus non-ADHD participants was < 85% for any one of the tests separately. The procedure significantly improved the score separation between ADHD versus non-ADHD groups. The final average probabilities for ADHD were 76% for the ADHD group and 8% for the control group. These probabilities correlated (r = .87) with the Brown ADD scale and (r = .84) with the ADHD-Symptom Inventory used for the screening of the participants. We conclude that, although each separate test was not completely accurate, a combination of several tests classified correctly all ADHD and all non-ADHD participants. The application of the proposed assessment is not limited to the specific tests used in this study--the assessment represents a general paradigm capable of accommodating a variety of ADHD tests into a single diagnostic assessment.

The EEG Consistency Index as a Psycho-Physiological Marker of ADHD and Methylphenidate Response: Replication of Results

“…Six electrode sites were prepared: a ground in front of Cz, a right earlobe reference electrode, and Cz, Pz, P3, P4. These sites were chosen in accordance with current EEG data research indicating positive findings with these sites (Robeva et al, 2004;Penberthy et al, 2005). EEG signals were amplified and processed by the Lexicor Neurosearch-24 system.…”

“…Although several studies have examined the EEGs of children while performing various tasks (e.g., reading and arithmetic), only one research group has examined the children specifically while they transition from one cognitive task to another cognitive task (Cox, Kovatchev et al, 1998;Cox, Merkel et al, 2000;Kovatchev et al, 2001;Robeva, Penberthy, Loboschefski, Cox, & Kovatchev, 2004;Penberthy et al, 2005). Because both clinical experience and research (Schachar, Tannock, Marriott, & Logan, 1993) support the notion that children with ADHD have difficulty cognitively transitioning from one task to another, Cox and his colleagues (Cox et al, 1998) measured differences between a small sample of boys with ADHD (n = 4) and non-ADHD boys (n = 4) as they cognitively shifted from watching a video to reading a book.…”

Calibration of ADHD Assessments Across Studies: A Meta-Analysis Tool

“…Subjects were administered the ADHD-SI, and the Wender-Utah Rating Scale (WURS; Ward, Wender, & Reimherr, 1993), which is a 61-item retrospective self-report scale with adequate reliability and validity. Results of analyses of these initial subjects are reported in Robeva et al, 2004.…”

“…In our recent paper (Robeva, Penberthy, Loboschefski, Cox, & Kovatchev, 2004) we proposed a new psychophysiological assessment procedure for diagnosing ADHD, capable of combining various disparate ADHD markers into a powerful assessment tool whose classification power exceeds the classification power of its separate components. The formal framework of that combined assessment employs a Bayesian algorithm that allows for linking of disparate assessment instruments into one unified probability assessment.…”

“…For example, in Robeva et al (2004) we employed data from twelve female college students using both self-report and others' report of symptoms, EEG information when transitioning from one cognitive task to another, behavioral information, cognitive performance, and history of symptoms. Classification of subjects as ADHD or non-ADHD improved with the sequential steps, achieving 100% accuracy by the final EEG measures.…”