Jin Hyung Lee: How can we systematically cure brain diseases?

Episode Summary

Executive Summary: Stanford’s Jin Li argues that brain medicine needs an engineering revolution: define normal brain function quantitatively, measure disease precisely, then use targeted interventions to restore function. The discussion covers why current diagnosis is often exclusionary, how multimodal measurement and computational modeling can reveal circuit-level dysfunction, and why future treatments may combine drugs, stimulation, gene editing, and stem cells with repeated feedback measurements.

Main Topics: Why brain disease needs an engineering approach (Priority: 5/5): Li contrasts biology’s slow, hypothesis-driven progress with engineering’s goal-setting and execution model, arguing brain disorders need systematic, measurable targets rather than vague descriptions. Limits of current diagnosis and treatment (Priority: 5/5): Current clinical practice often relies on exclusion and symptoms because tools rarely identify the underlying algorithmic dysfunction in diseases like Alzheimer’s, Parkinson’s, epilepsy, and stroke. Measurement as the foundation for intervention (Priority: 5/5): MRI, CT, EEG, functional imaging, and other tools are useful but insufficient; the needed leap is to quantify brain status well enough to define normal versus disease states. Multi-resolution brain modeling (Priority: 4/5): Li’s lab combines genetic tools, imaging, and electrical recordings to reconstruct brain circuits from large-scale networks down to cell-type and neuron-level interactions. AI as an aid, not the core model (Priority: 4/5): Deep learning and related methods are used mainly to clean noisy experimental data and clarify signals; the actual models are grounded in direct measurements and behavior. Future intervention toolkit (Priority: 5/5): Potential treatments include drugs, brain stimulation, gene editing, and stem cells, but they must be deployed with precise measurement and repeated monitoring to avoid overcorrection. Disease-specific progress and clinical readiness (Priority: 4/5): Epilepsy and Parkinson’s are closer to actionable models, while Alzheimer’s remains a major challenge but is increasingly being decomposed into measurable cognitive circuit problems.

Key Arguments: Brain disorders remain hard to treat because medicine lacks a systematic, engineering-style way to define goals, measure function, and optimize interventions. Hypothesis-driven biology can explain isolated findings, but it does not by itself tell clinicians what to change or how to restore normal function. If clinicians could measure a patient’s brain state against a defined normal range, treatment could move from trial-and-error to targeted correction. The brain should be viewed as an algorithmic system with circuit-level functions that can be reconstructed from multimodal data. Current tools are useful but not enough to diagnose disease mechanisms directly; most neurological diagnoses are still based on exclusion. AI should support signal extraction and data cleanup, but the main scientific model must come from experimentally grounded measurements of real brain behavior. Effective brain therapy will likely require multiple knobs—combining drugs, stimulation, gene editing, and possibly stem cells—rather than a single cure. Repeated post-treatment measurement will be essential to determine whether an intervention has moved the patient toward normal function or overshot. Epilepsy and some Parkinsonian motor-circuit problems are becoming sufficiently understood to approach clinical translation, while Alzheimer’s requires more work to map cognitive dysfunction precisely.

Data Points: NIH Brain Project duration: 10-year project - Referenced as a major national effort to advance brain science and measurement tools. Number of measurement modalities discussed: At least 5 - CT, MRI, EEG, functional imaging (glucose/oxygen use), genetic tools, and electrical recordings were all mentioned. Clinical diagnosis approach: Diagnosis of exclusion - Described as the common approach for neurological diseases when direct markers are unavailable. Brain function target: Normal range of function - Li emphasizes defining a measurable normal state to guide therapy. Therapy optimization: Multiple knobs - Used to describe combining interventions such as drugs, stimulation, and gene editing.

Pivotal Quotes: "if you can't measure, you can't intervene." — Jin Li: Li’s core engineering principle for treating brain disease. "What you need is to be able to define the normal function." — Jin Li: Explaining the treatment goal for disorders like Parkinson’s and Alzheimer’s. "the more knobs you have, the better." — Jin Li: Discussing why future treatment may combine drugs, stimulation, gene editing, and stem cells.

Implications: The conversation points toward precision neurology: objective brain metrics, circuit-level modeling, and feedback-guided multi-modal therapies. If successful, diagnosis and treatment could shift from guesswork to personalized, measurable control of brain function.