Understanding & Healing the Mind | Dr. Karl Deisseroth

Unlike neurology, which has scans and EEGs that visibly reveal strokes or seizures, psychiatry has no blood test or imaging marker that can diagnose depression, schizophrenia, or autism in an individual. Diagnoses rely on patients’ words, behavior, and rating scales, which are all linguistic and observational rather than direct physical measures. This makes interviewing an art as well as a science: good psychiatrists must translate vague or colloquial terms like “I’m depressed” into precise, observable features (e.g., inability to imagine tomorrow, hopelessness, early-morning awakening, changes in appetite or movement).

Cognitive behavioral therapy can largely resolve panic disorder in 6–12 structured sessions for motivated patients. Antipsychotic medications, especially clozapine, can dramatically reduce hallucinations and paranoia in schizophrenia, albeit with substantial side effects. Electroconvulsive therapy (ECT) is extraordinarily effective for treatment-resistant depression but functions via a brain-wide seizure with poorly understood mechanisms. Vagus nerve stimulation and deep brain stimulation help subsets of patients, yet they act broadly on mixed fibers because electrical stimulation activates everything near the electrode.

Channelrhodopsins originated as light-sensing proteins in single-celled algae that swim to optimal light levels. Deisseroth’s lab repurposed their genes into neurons using viral vectors (AAVs), then used light to turn specific cell types on or off in behaving animals, revealing which circuits control particular actions and states. Beyond direct human applications (e.g., partial vision restoration in a blind patient), Deisseroth emphasizes that the biggest clinical impact will be “indirect”: mapping exactly which cells and projections drive symptoms like anhedonia or passive coping, then using that knowledge to design drugs that selectively target those cells’ unique receptors rather than blanket neurotransmitter systems.

Deisseroth’s group showed that dissociation—a separation of self from bodily experience common in trauma and certain drugs like ketamine—can be modeled in mice and linked to specific activity patterns in a conserved brain region. They recorded similar patterns in a human epilepsy patient who experienced dissociation before seizures, and optogenetically reproduced dissociative-like behavior in mice by replaying that pattern. For depression, they study circuits involving dopamine and active vs. passive coping, such as interactions between habenula and raphe. For autism, they frame social interaction as an extreme information-integration problem, guiding them to study circuits that fuse complex, high–bit-rate data streams (voice, eye gaze, movement) in social animals.

Deisseroth views invasive tools like closed-loop deep brain stimulation and brain–machine interfaces as part of psychiatry’s future, especially for severe, refractory conditions. However, he believes the major advance will be medications and noninvasive treatments grounded in optogenetically-derived circuit maps—knowing which cell type from point A to point B actually matters for a given symptom. Once those causal nodes are known, their receptor expression and connectivity can guide the design or repurposing of drugs, potentially yielding fast paths to clinical trials using already-approved compounds.