The CMET is founded on the key idea that future clinical diagnoses are likely to benefit from quantitative indices (physiomarkers) constructed by use of data-based predictive dynamic models of the subject-specific physiological processes underlying a disease.
An example is the use of model-based indices of cerebral hemodynamics (flow autoregulation and vasomotor reactivity) for improved diagnosis of Alzheimer’s disease or other neurodegenerative diseases with vascular components (such as vascular dementia, mild cognitive impairment etc). The obtained data-based model describes the dynamic nonlinear relationship between arterial blood pressure, cerebral blood flow and end-tidal CO2 (all variables measured as time-series data non-invasively, safely and inexpensively in a clinical context). The model is constructed by use of our pioneering methodologies that have been developed over the years within the context of our Biomedical Simulations Resource (BMSR) — a research center founded by Prof. Marmarelis in 1985 that has received uninterrupted peer-reviewed funding by the National Institutes of Health. The diagnostic efficacy of the particular physiomarker of cerebral vasomotor reactivity is being examined in our current work and has been shown to yield excellent diagnostic results for early-stage Alzheimer’s. Confirmation of this capability in a large population of patients will provide the means for improved diagnosis and treatment monitoring – translating into better care for millions, greater patient comfort and huge annual savings for our health care system.
Another example (see figure below) is in the area of neurostimulation for the treatment of intractable neuropsychiatric disorders (e.g. major depression or post-traumatic stress disorder). Our group has started pioneering research in developing the “next-generation” closed-loop Deep Brain Stimulation (DBS) system that will be capable of treating these debilitating disorders by applying optimal spatio-temporal stimulation patterns at various brain regions (e.g. the Nucleus Accumbens and the Amygdala shown in the figure). These optimal patterns will be designed on the basis of sophisticated dynamic nonlinear models of the relevant multi-neuronal interactions (see below the illustrative block-diagram of a PDM-based model of the inter-connectivity in the pre-frontal cortex obtained in pilot studies via our novel methodologies). These models can be computed on-line from direct recordings of neuronal activity. This will endow the envisioned next-gen DBS system with on-line adaptive capability to identify the presence of an event related to the disorder and apply the optimal neurostimulation pattern to treat it in a closed-loop manner.