Contrast Processing Deficits in Schizophrenia and Schizoaffective Disorder: Contributions to Neurocognition, Facial Emotion Recognition, and Functional Outcome

Abstract

Background: Individuals with schizophrenia exhibit visual dysfunction. This study investigated contributions of contrast processes to higher-level functions in schizophrenia. Study 1 measured contrast processing using electrophysiological and psychophysical measures and tested deficits as a classification tool for schizophrenia. Relationships were explored between contrast processing, neurocognition, emotion recognition, and functional outcome in patients. Study 2 used a larger sample size to replicate psychophysical findings and further explore relationships. Methods: Participants were patients with schizophrenia and schizoaffective disorder (Study 1 n = 40; Study 2 n = 93) and age-matched healthy controls (Study 1 n = 39; Study 2 n = 93). Steady-state visual evoked potentials (ssVEPs) elicited by hexagonal arrays of dots (contrast sweep; 1-32% contrast) provided measures of signal-to-noise ratio (SNR), amplitude, phase, and noise. A biophysical model applied to ssVEP contrast-response functions estimated contrast gain (response amplification at low contrast) and gain control (response compression and phase advance with increasing contrast). A two-alternative spatial forced-choice paradigm (horizontal gratings; 0.5-21 cycles/degree) measured psychophysical contrast sensitivity (CS). Multilevel linear modeling examined group differences in ssVEP and CS measures. Predictive accuracy for group classification was estimated through receiver-operating-characteristic (ROC) curve analyses. Neurocognition (MCCB/WAIS-III), facial emotion recognition (ER-40), and functional outcome (UPSA) were assessed. Results: Patients showed reduced contrast gain (Mpatients = 4.50 muv, SD patients = 3.38; Mcontrols = 8.35 muv, SDcontrols = 5.68,p < .01), marginally reduced gain control (Mpatients = 7.53 muv, SDpatients = 6.82, Mcontrols = 10.88 muv, SDcontrols = 5.92, p = .07), and reduced CS (all p < .001). Patients showed lower SNRs than controls (F(1,69.69) = 9.36,p = .003) due to reduced amplitudes rather than greater noise. A single variate from discriminant analysis of ssVEP and CS responses produced maximal separation (AUC = .94, 95% CI [0.88, 1.00]). Low spatial frequency CS (0.5, 1 cycles/deg) mediated the relationship between SNR and neurocognition, and moderated the relationship between emotion recognition and functional outcome. Conclusions: Results support impaired contrast processing in schizophrenia and provide evidence for visual deficits as a biomarker. Further characterization of the role of impaired contrast detection in cognition and functioning may inform treatments.