Developing Precision Invasive Neuromodulation for Psychiatry

Abstract

Psychiatric conditions are common and often disabling. Although great strides have been made in alleviating symptoms with pharmacotherapy and psychotherapeutic approaches, many patients continue to have severe disease burden despite the best therapies available. One of the pervasive challenges to improving treatment is that present diagnostic and therapeutic strategies lag behind our modern conceptualization of the pathophysiology of these disorders. Psychiatric symptoms manifest through activity in specific neural circuits; thus, therapies capable of modulating these circuits are attractive. The investigators reviewed recent advances that facilitate treating medically refractory psychiatric disorders with intracranial neuromodulation in a way that intervenes more directly with the underlying pathophysiology. Specifically, they reviewed the prospects for using intracranial multielectrode arrays to record brain activity with high spatiotemporal resolution and identify circuit-level electrophysiological correlates of symptoms. A causal relationship of circuit electrophysiology to symptoms could then be established by modulating the circuits to disrupt the symptoms. Personalized therapeutic neuromodulation strategies can then proceed in a rational manner with stimulation protocols informed by the underlying circuit-based pathophysiology of the most bothersome symptoms. This strategy would enhance current methods in neurotherapeutics by identifying individualized anatomical targets with symptom-specific precision, circumventing many of the limitations inherent in modern psychiatric nosology and treatment.

Keywords: Deep Brain Stimulation; Epilepsy; Mood Disorders; Neuromodulation; Neurostimulation; Psychiatry.

FIGURE 1.

Invasive neuromodulation treatment outline^{a a} With major depressive disorder as an example, patients with a treatment-resistant course would have electrode arrays implanted intracranially in predetermined regions relevant for current symptoms. They would undergo symptom monitoring and quantification while data from electrodes were continuously recorded. Data would be annotated by the patient when specific intrusive thoughts or worsening symptoms occurred. Neural decoding would occur via machine-learning algorithms to correlate electrophysiological data with ongoing symptoms. Stimulation would then be delivered; if symptoms improved, it would be inferred that a casual neural circuit was identified. Therapeutic brain stimulation could then be delivered in a specific circuit-driven way. The closed-loop nature of this would allow for cycling back to mapping of causal circuitry, thereby improving the stimulation delivered as symptoms change.