Improving causality perception judgments in schizophrenia spectrum disorder via transcranial direct current stimulation

Abstract

Background: Deficient causality perception and attribution may underlie key symptoms of schizophrenia spectrum disorder (SSD), such as delusions and ideas of reference. Although transcranial direct current stimulation (tDCS) can increase the influence of spatial information on perceptual causality judgments among healthy participants, its effect among patients with SSD remains unknown. We sought to determine whether tDCS modulates the contribution of stimulus characteristics to perceptual causality judgments among patients with SSD; we predicted that right parietal tDCS would increase the influence of spatial stimulus characteristics on patients' causality perception.

Methods: Patients with SSD received frontal, parietal, frontoparietal and sham tDCS in 4 separate sessions. Pre- and post-tDCS, patients viewed video clips of ball A colliding with ball B. Spatial linearity (ball B's angle of egress) and temporal contiguity (delay between collision and ball B's movement) varied parametrically. After each launching event, patients rated perceived causality.

Results: Among 19 patients with SSD, we found a brain region-dependent effect of tDCS regarding sensitivity to violations of spatial linearity. After right parietal anodal tDCS, the influence of angle variations on patients' perceptual causality judgments increased, reflected by a higher probability of perceived causality for stimuli with small angles and a lower probability of perceived causality for stimuli with high angles.

Conclusion: Transcranial direct current stimulation increased the influence of spatial stimulus characteristics on causality perception among patients with SSD. Future research should explore potential links between tDCS-induced changes in basic perceptual processes and clinical symptoms, such as delusions and ideas of reference.

Figure 1

(A) Transcranial direct current stimulation (tDCS) conditions, showing the simulated normalized electric field strengths (in V/m) on the cortical surface for the 3 different tDCS conditions (from left to right: parietal [left parietal cathodal, right parietal anodal], frontoparietal [left frontal cathodal, right parietal anodal] and frontal [left frontal cathodal, right frontal anodal] tDCS). In addition, every patient received 1 session of sham tDCS. (B) Procedure and task. In each session, before and after 10 minutes of 1.5 mA tDCS, 98 stimuli of ball A launching ball B were presented on screen. Each of the 98 possible combinations of stimulus characteristics (7 different angles, 7 different delays, presented right to left and left to right) was presented exactly once, in pseudorandomized order. Immediately after each stimulus, patients judged perceived causality by pressing a button, either yes (causal) or no (noncausal).

Figure 2

Location-specific effect of transcranial direct current stimulation (tDCS) on angle sensitivity of perceptual causality judgments. (A) The estimated marginal means for the probability of causal perception (i.e., causal judgments) for different angles of egress of the second ball (measured in degrees), for the 4 different tDCS locations, pre- and post-tDCS. Error bars indicate standard errors. (B) Location by angle interaction plots of estimated marginal means (i.e., the linear prediction of the generalized linear mixed models for causal perception based on angle (mean-centred), pre-tDCS and (C) post-tDCS, for the 4 different tDCS locations. LFC-RFA: left frontal cathodal, right frontal anodal tDCS; LFC-RPA: left frontal cathodal, right frontal anodal tDCS; LPC-RPA: left parietal cathodal, right parietal anodal tDCS.

Figure 3

Angle sensitivity of causality judgments after left frontal cathodal, right parietal anodal (LFC-RPA) transcranial direct current stimulation (tDCS); shown are the estimated marginal means for the probability of causal perception (i.e., causal judgments) for different angles of egress of the second ball, pre- and post-tDCS. False discovery rate–adjusted p values were computed using the Benjamini–Hochberg method, to correct for performing 28 tests (Table 2). Error bars indicate standard errors. *p_adjusted < 0.05.