USING TRANSCRANIAL MAGNETIC STIMULATION TO EXAMINE AND MODULATE BRAIN CIRCUITS IN SCHIZOPHRENIA

Schizophrenia is a highly heterogeneous disorder that affects up to 1% of the world's population and ranks among the leading causes of disability worldwide. The clinical picture is characterized by chronic psychotic symptoms, episodic relapses, and significant disruption of social functioning. Symptoms are often divided into “positive”, including hallucinations and delusions, and “negative”, such as social withdrawal and anhedonia. In most patients, these are associated with a gradual deterioration in cognitive functions. Of note, while current treatments based on antipsychotic medications are effective at treating positive symptoms and, to a lesser extent, negative symptoms, there is a dearth of available interventions able to ameliorate the cognitive deficits associated with schizophrenia. Furthermore, the causes and mechanisms underlying this condition remain elusive. Accumulating evidence suggests that abnormalities in the ability of neurons to synchronize their oscillatory activity, particularly in the beta and gamma frequency bands, might be implicated in the pathophysiology of schizophrenia. Evidence from studies employing electroencephalography (EEG) across several paradigms (resting state, steady state, sensory event-related, task-related, etc.) suggests that deficits in these “fast” brain rhythms in the frontal lobe are common in schizophrenia and might be associated with its clinical symptoms, particularly cognitive dysfunction. These alterations are believed to reflect an imbalance between excitatory and inhibitory function, particularly related to a dysfunction of specific populations of cortical and subcortical GABAergic interneurons. However, conventional EEG studies, particularly when investigating task-related neural activity, can be affected by the subject’s engagement, attention, motivation, or clinical symptoms. This represents an inherent limitation to their application for the development of useful biomarkers in severe psychiatric disorders such as schizophrenia. Furthermore, a large gap remains between the abundance of evidence on altered fast oscillations in schizophrenia and the lack of effective interventional strategies developed around these putative biological targets. Transcranial magnetic stimulation (TMS) is a neurostimulation technique that leverages electromagnetic induction to non-invasively induce targeted electric currents in the brain cortex. One important advantage of TMS over other brain stimulation techniques is that it can be employed as both an exploratory technique, when coupling single pulses of TMS with EEG recordings (TMS-EEG), and as an interventional tool, when TMS is used in repetitive patterns to induce brain plasticity (rTMS). TMS-EEG allows for non-invasively probing brain areas and circuits and extracting quantitative measures of excitability, oscillatory function, and connectivity. Importantly, the perturb-and-measure nature of TMS-EEG makes it relatively independent from subject-specific behavioral confounders, such as attention, motivation, and cognitive abilities, making it an ideal tool to assess the integrity (or lack thereof) of brain circuits in neuropsychiatric populations. One prominent TMS-EEG measure of (thalamo)cortical function is the main oscillatory EEG frequency recorded after the TMS pulse, or “natural frequency”. In the healthy brain, different brain areas respond with different natural frequencies: the prefrontal natural frequency is in the gamma (30-40 Hz) frequency band, the premotor natural frequency is in the high beta/low gamma (25-34 Hz) range, the parietal natural frequency is in the beta-1 (13-25 Hz) band, and the occipital natural frequency is in the alpha (8-12 Hz) range. This rostrocaudal gradient likely reflects differences in the intrinsic oscillatory properties of these brain areas, which may underlie their functional specificities. Recent evidence from TMS-EEG has revealed that, in chronic schizophrenia, brain areas of the frontal lobe - including motor, premotor, and prefrontal areas - show alterations in several TMS-evoked EEG parameters, compared to healthy control (HC) subjects. In particular, one study found that the natural frequency is significantly reduced in these areas in chronic schizophrenia, compared to HCs, with the prefrontal natural frequency showing a complete separation between patients and healthy subjects. Furthermore, in schizophrenia patients, the prefrontal natural frequency correlated negatively with clinical symptoms and working memory performance. However, it is unclear whether these deficits are related to confounding factors linked to the chronicity of the illness, such as length of exposure to antipsychotic medications, comorbidity, or neurodegenerative processes, or rather are present since early in the course of schizophrenia, thus being more likely linked to core pathophysiological mechanisms associated with this disorder. When TMS pulses are delivered repetitively, as in rTMS, they can induce neuronal plasticity and modify local excitability and long-distance connectivity. rTMS protocols include “excitatory” high-frequency stimulation (where pulses are delivered at a frequency of 5-10 Hz), “inhibitory” low-frequency stimulation (1 Hz), and novel theta burst stimulation (TBS) paradigms. In TBS, 50Hz triplets of TMS pulses are delivered at 5-8 Hz (i.e., theta frequency) in either intermittent blocks (iTBS, excitatory) or continuously (cTBS, inhibitory). Administering multiple sessions of rTMS (over 3-4 weeks) is a currently approved treatment for medication-resistant depression, obsessive-compulsive disorder, and migraine, as well as for smoking cessation. However, its potential role in ameliorating oscillatory dysfunction and cognitive deficits associated with schizophrenia is still underexplored. In this PhD thesis, I will first report on three studies that were recently conducted at Prof. Fabio Ferrarelli’s lab at the University of Pittsburgh. In these studies, we employed TMS-EEG to explore the oscillatory properties of different cortical areas and their association with clinical symptoms in patients in the early stages of schizophrenia. The overarching goal of these studies was to test the role of TMS-EEG measures, including the natural frequency, as candidate pathophysiological biomarkers of schizophrenia. I will then outline the study design of a recently funded mechanistic clinical trial that leverages findings from study 3 to test the efficacy of iTBS in acutely ameliorating TMS-EEG parameters and cognitive performance in subjects with schizophrenia. Preliminary results from an initial subset of subjects will be briefly discussed. In study 1, we used TMS-EEG to probe the left motor and left posterior parietal areas in 23 acutely ill, first-episode psychosis (FEP) patients and 13 age and gender-matched HC subjects. Patients were assessed with clinical scales by experienced raters at baseline and during a six-month follow-up. From TMS-EEG data, we calculated the natural frequency as well as the EEG spectral power in each frequency band relative to the broadband power, or Relative Spectral Power (RSP). The natural frequency did not show any significant group difference in either the motor or posterior parietal cortices. No significant differences were found in any other TMS-EEG parameter for the posterior parietal cortex. However, when stimulating the motor cortex, FEP patients showed a local deficit in the high beta/low gamma RSP, relative to HC participants, corresponding to a large effect size (Cohen’s d = 1.5). High beta/low gamma deficits correlated with worse positive symptoms at baseline and predicted their severity at the six-month assessment. Given its association with positive symptoms in acutely ill patients, this reduction may represent a state biomarker of acute psychosis. In study 2, we used TMS-EEG to target the left premotor cortex in 16 outpatients with early-course schizophrenia patients (<2 years from onset) and 16 age and gender-matched HCs and found that the natural frequency of the premotor cortex was significantly reduced in early-course schizophrenia, corresponding to a large effect size (Cohen’s d = 1.06). No correlation was found between the natural frequency and age, clinical symptom severity, or dose of antipsychotic medications. In study 3, we applied TMS-EEG to the left dorsolateral prefrontal cortex (DLPFC) in 30 outpatients with early-course schizophrenia and 28 age and gender-matched HCs. Goal-directed working memory performance was assessed using the “AX” Continuous Performance Task (AX-CPT). The natural frequency of the left DLPFC was significantly reduced in patients compared to HCs, which corresponded to a very large effect size (Cohen’s d > 2.0). Patients also showed a higher local RSP in the beta band, relative to HCs. The AX-CPT performance was significantly worse in patients relative to HCs. Across all participants, the beta-band RSP in the left frontal EEG channels correlated inversely with the AX-CPT performance. Altogether, these findings indicate that intrinsic oscillatory deficits are present in frontal areas since the early stages of schizophrenia and may relate to its clinical symptoms, supporting their role as candidate pathophysiological biomarkers of this disorder. Interestingly, these EEG deficits exhibit a postero-anterior progression similar to what is seen in chronic patients, with prefrontal areas showing the largest impairment. Particularly, results from study 3 suggest that a reduction in the natural frequency of the DLPFC may represent an early neural signature associated with working memory deficits often seen in schizophrenia, supporting the existence of a link between prefrontal fast oscillatory dysfunction and cognitive impairment in schizophrenia. These findings may indicate a potential avenue for intervention to address a historically unmet need in these patients. Finally, we report the study design and preliminary findings of an interventional double-blind, placebo-controlled, randomized, cross-over mechanistic clinical trial in a cohort of early-course schizophrenia patients (current n=20; projected recruitment n=75). Building on findings from study 3, this ongoing study combines TMS-EEG and rTMS to test the effectiveness of a single session of iTBS of the DLPFC to acutely a) revert the abnormalities in prefrontal TMS-evoked oscillatory EEG activity, including the reduction of the DLPFC natural frequency; and b) improve the working memory performance.