Future clinical uses of neurophysiological biomarkers to predict and monitor treatment response for schizophrenia

Abstract

Advances in psychiatric neuroscience have transformed our understanding of impaired and spared brain functions in psychotic illnesses. Despite substantial progress, few (if any) laboratory tests have graduated to clinics to inform diagnoses, guide treatments, and monitor treatment response. Providers must rely on careful behavioral observation and interview techniques to make inferences about patients' inner experiences and then secondary deductions about impacted neural systems. Development of more effective treatments has also been hindered by a lack of translational quantitative biomarkers that can span the brain-behavior treatment knowledge gap. Here, we describe an example of a simple, low-cost, and translatable electroencephalography (EEG) measure that offers promise for improving our understanding and treatment of psychotic illnesses: mismatch negativity (MMN). MMN is sensitive to and/or predicts response to some pharmacologic and nonpharmacologic interventions and accounts for substantial portions of variance in clinical, cognitive, and psychosocial functioning in schizophrenia (SZ). This measure has recently been validated for use in large-scale multisite clinical studies of SZ. Finally, MMN greatly improves our ability to forecast which individuals at high clinical risk actually develop a psychotic illness. These attributes suggest that MMN can contribute to personalized biomarker-guided treatment strategies aimed at ameliorating or even preventing the onset of psychosis.

Keywords: biomarker; cognitive remediation; mismatch negativity; neurocognition; schizophrenia.

Figure 1

Effect size (Cohen’s d) of deficits in schizophrenia patients across leading candidate biomarkers. Data from Ref. . Abbreviations: LNS, letter number sequencing; WCST, Wisconsin Card Sorting Test; CVLT-II, California Verbal Learning Test—second edition; PPI, prepulse inhibition.

Figure 2

Example of overlapping distributions in a robust (d = 1) effect size biomarker deficit in schizophrenia patients. In neuropsychological assessments, d = 1 standard deviations below the mean is commonly used for impairment classification. With an effect size d = 1 below the mean, 50% of patients exhibit unimpaired/normal-range biomarker values. Data from Ref. .

Figure 3

MMN/P3a paradigm and group averages. Participants were presented with stimuli consisting of frequently presented standard stimuli (90% of trials, red box labeled “s”) interspersed with infrequent deviant stimuli (10% of trials, blue box labeled “deviant”). ERP waves to standard and deviant stimuli are calculated by averaging EEG responses to each stimulus type. Deviant–standard difference waves are generated by calculating MMN and P3a components (black lines). For all waveforms, solid lines represent healthy comparison subjects (n = 753) and dotted lines are used for schizophrenia patients (n = 877). From Ref. .

Figure 4

1-year stability of neurophysiological and neurocognitive biomarkers. Intraclass correlation coefficients are shown for schizophrenia patients (blue; n = 163) and nonpsychiatric comparison subjects (red, n = 58). The mean retest interval was 364.57 (SD: 23.83) days. Data from Ref. . Abbreviations: LNS, letter number sequencing; WCST, Wisconsin Card Sorting Test; CVLT-II, California Verbal Learning Test—second edition; PPI, prepulse inhibition.

Figure 5

Individual subject and group averaged waveforms. Individual subject deviant–standard difference wave averages (color coded by amplitude) are shown in the top panel for healthy comparison subjects (n = 753) and schizophrenia patients (n = 877). Group grand average wave forms are shown in the bottom panel. Data from Ref. .

Figure 6

Future clinical use of laboratory-based biomarkers to assign patients to treatments.