Extended Accelerated tDCS and correction of prediction error signalling in Schizophrenia with atypical hallucinations: A Case Report

Dear Editor,

Meta-analyses suggest clinical benefit from tDCS treatment for auditory hallucinations (AH) in schizophrenia (SCZ) (Guttesen et al., 2021; Rashidi et al., 2021), but the potential clinical utility of tDCS treatment in SCZ with prominent basic symptoms of self-awareness is unknown. Recent reports have explored the safety, tolerability and clinical utility of accelerated tDCS (more sessions over a shorter duration) in schizophrenia (Mondino et al., 2021) and associated neurophysiological changes relevant to AH pathophysiology (Nayok et al., 2022).

Capturing changes in event-related potential (ERP) in response to therapeutic tDCS may elucidate treatment response (Wang et al., 2020). Roving mismatch negativity (rMMN) task encapsulates components of prediction error signalling and reflects short-term plasticity (McCleery et al., 2019). In rMMN, the deviant becomes the new standard over the course of stimuli presentation. The strength of prediction error signalling elicited by the deviant stimuli varies in response to the standard stimuli’s train preceding it. rMMN paradigm examines: repetition positivity–the strength of standard stimulus’s memory trace that increases with repetition; deviant negativity−the strength of prediction error signalling; and mismatch negativity–the magnitude of elicited prediction error (McCleery et al., 2019).

In this case report, we explored the safety, utility and feasibility of extended, accelerated tDCS therapy in a schizophrenia patient with atypical, recalcitrant symptoms. We also examined concurrent changes in short-term plasticity using a rMMN paradigm.

Case Report

Mr F is a 29 years old male patient with a history of nine years of schizophrenia (DSM-5) (Association, 2013). He had 1) delusion of misinterpretation of people’s gestures and actions; 2) reflex hallucination whereby any random auditory stimuli like a blowing horn or chirping of birds in the immediate environment elicited itching in his ears; such stimuli were also misinterpreted as messages or communication aimed at him; and 3) cenesthetic hallucinations of being touched or caressed when someone stood nearby. Examination of Anomalous Self-Experience (EASE) interview at baseline revealed severe self-referential experiences, moderate severity of captivation of attention by details in the perceptual field, cenesthetic experiences, social anxiety and threatening bodily contact with unpleasant feelings, and mild disturbance in experience of time. Psycho-Sensory hallucinations Scale (PSAS) indicated significant cenesthetic hallucinations (20/23). Clinical Global Impression severity (CGI-S) Scale had a rating of 5 (that is, markedly ill).

The patient presented with a partial response to clozapine (300 mg/day). Add-on accelerated left fronto-temoporo-parietal 2mA tDCS was administered (2-days, 5-sessions/day with 20-minute intersession intervals) (Mondino et al., 2021). Side-effects were systematically monitored (Brunoni et al., 2011). As the patient tolerated the sessions well and indicated slight improvement, another course (2 days, 10 sessions) of accelerated tDCS (acctDCS) was administered. Short-term auditory plasticity was examined at three time-points—pre acctDCS, post acctDCS-1^st course and post acctDCS-2^nd course—using rMMN ERP paradigm (McCleery et al., 2019). Clinical outcome was assessed using PSAS and patient’s self-rating on an 11-point visual analogue scale of increasing severity. Clinical Global Impression Scale-Improvement (CGI-I) rating was also taken. Brief neuromodulation is unlikely to alter features of psychopathology like nature of anomalous self-related subjective experiences; therefore, EASE scale was not considered as an outcome measure.

Transient improvement in clinical symptoms was noted on the patient-rated visual analogue scale (Baseline=5; post acctDCS-1^st course=3; post acctDCS-2^nd course=5). PSAS score did not change. The CGI-I rating was 3, that is, minimally improved. However, the patient reported enhanced ability to cope with the symptoms—a prominent reduction in salience of symptoms—which was a clinically meaningful recovery for him and his caregiver. the patient did not report reduction in the symptom severity, however, despite the reflex AH and cenesthetic hallucinations, he could take a walk outside, hold a conversation with the caregiver and cooperate during assessments by tolerating the symptoms well; these things were extremely distressing prior to the treatment. This change underscores significance of coping—a critical component of prognosis—that is related both to vulnerability and resilience. A recent meta-analysis has linked lower resilience to poor psychosocial functioning in schizophrenia-spectrum disorders (Wambua et al., 2020). Maladaptive coping (like avoidance and withdrawal), as acknowledged by the patient, had adversely affected his activities of daily living, and severely restricted his scope for social connections.

The correction of rMMN profile suggests improvement in prediction error signalling (Figure 1). The abnormal rMMN profile at baseline likely reflects the aberrant sensory association between irrelevant stimuli and related internal misrepresentations. The gradual course correction of the rMMN profile is potentially due to the modulation of short-term auditory plasticity with acctDCS. The aberrant salience theory of psychosis posits that antipsychotics do not terminate abnormal experiences; instead, they “dampen the salience” of abnormal experiences and facilitate a platform for “a process of psychological resolution” (Kapur, 2003). Additionally, a paradoxical learning deficit has been well documented in schizophrenia, expressed as an overlearning from irrelevant or neutral stimuli and concomitant inadequate learning from reward-predicting information (Millard et al., 2022). AcctDCS seems to have addressed this excess salience towards irrelevant environmental cues (indicated by improved rMMN profile) that was leading to abnormal internal representations and associated anomalous self-experiences like reflex AH and cenesthetic hallucinations. Though the neuroplasticity induction with acctDCS and concurrent changes in psychopathology-related neurophysiology was rapid, the clinical outcome measures did not change in tandem. This observation suggests clinical outcomes may operate on a different timeline. Thus, clinical researchers may benefit from repeating clinical assessments at multiple follow-up time-points for better clinical acuity. These clinical assessment time-points may intentionally differ from the timeline of neurophysiological assessment complementing them.

Figure 1.

Roving MMN (rMMN) profile of the patient at pre-acctDCS, post-acctDCS-1st course and post-acctDCS-2nd course. rMMN is a nuanced measure of prediction error; contingent upon repetition positivity (RP) and deviant negativity (DN) profiles. The expected RP profile is supposed to be RP₃₃<RP₈<RP₃ in the order of increasing negativity of N1 amplitude; the greater the number of repetitions, the stronger the memory trace and the less negative/more positive the N1 amplitude. The reverse is true for DN and MMN; the expected pattern is DN₃₃>DN₈>DN₃ and MMN₃₃>MMN₈>MMN₃ in the order of decreasing negativity of N1 amplitude. The magnitude of prediction error signalling is higher when the train of preceding standard stimuli is longer.

At baseline/pre-acctDCS, the rMMN profile of the patient is notably impaired [Figure 1(a)]. MMN₃₃ amplitude is notably suppressed, indicating a significant deficiency in prediction error. Post acctDCS 1^st course, a discernible improvement in the DN and MMN profiles can be noted, though the RP profile is still disrupted [Figure 1(b)]. Post acctDCS 2^nd course, a definitive correction in the RP profile alongside partial correction of DN and MMN profiles [Figure 1(c)] is evident. Though the ideal pattern of DN₃₃>DN₈>DN₃ and MMN₃₃>MMN₈>MMN₃ in the order of decreasing negativity of N1 amplitude is not achieved, the DN₃₃ and MMN₃₃ amplitudes are no longer attenuated as they were at baseline, indicating enhancement of prediction error signalling.

The present case study supports the safety, tolerability, and potential clinical utility of extended acctDCS (20-sessions/four-days) for treating atypical hallucinations and complex symptoms in schizophrenia. Due to lower prevalence, atypical hallucinations are under-researched; this also limits the scope of this case report. Correction of rMMN profile suggests acctDCS a) is capable of rapid neuroplasticity modulation, b) corrects deficient prediction error signalling and dampens excessive salience in schizophrenia. These features may have improved information processing, reduced internal misrepresentations of environmental stimuli paving the way to enhanced coping in the patient.

AcctDCS protocol could be a suitable alternative for treating atypical, intractable symptoms of psychosis. Meaningful neurophysiological assessment may complement standard questionnaires to substantiate clinical assessment and illuminate the neural underpinnings of improvement. Whether clinical benefit and potential for neuroplasticity modulation varies between acctDCS and standard tDCS, future studies should examine.