Neurogenis Foundation Lab

Building the first whole-brain foundation model trained on high-density Optically Pumped Magnetoencephalography (OPM-MEG) sensor streams.

Core Mission: The Brain Foundation Model

Human cognition operates at millisecond latency across macroscopic re-entrant loops. Modern AI models train on symbolic human outputs, while neuroimaging has been limited to sluggish hemodynamic blood-oxygen responses (fMRI) or blurred scalp voltages (EEG).

Neurogenis collects continuous whole-head magnetic recordings at sub-femtotesla sensitivity and 1,000 Hz using quantum optically pumped magnetometers. By pretraining self-supervised causal transformers directly over these electrodynamic fields, we are constructing a foundation model of the living human brain.

Epistemic Map: What Neuroscience Knows vs. Does Not Know

Our interactive 3D model encodes scientific understanding through light and material only. Luminous structured regions represent characterized microcircuits; soft, fogged, shimmering regions represent our lab's active scientific frontiers:

Prefrontal–Mediodorsal Thalamocortical Loop (Epistemic Score: 0.14 — Primary Unknown): How re-entrant sub-second thalamocortical phase coupling coordinates credit assignment and high-level reasoning.
Claustro-Insular Salience Complex (Epistemic Score: 0.18): The biophysical routing mechanism for binding multimodal sensory signals into unified conscious awareness.
Default Mode & Precuneus Narrative Complex (Epistemic Score: 0.24): The electrodynamic syntax of internal counterfactual simulation and autobiographical memory.
Hippocampal–Entorhinal Complex (Epistemic Score: 0.42): Mechanisms transforming continuous experience into discrete memory tokens.
Perisylvian Language Network (Epistemic Score: 0.54): Sub-10ms compositional syntax binding words into semantic hierarchies.
Cerebellar Predictive Forward Microzones (Epistemic Score: 0.68): Purkinje forward-model computation scaling from motor primitives to cognitive error signals.
Primary Sensorimotor Rolandic Axis (Epistemic Score: 0.82): Well-mapped Penfield homunculus and motor vector coding.
Early Calcarine Visual Cortex (Epistemic Score: 0.88): Deeply understood orientation columns and receptive field hierarchies.

Founders & Executive Leadership

Samer Zumot — Co-Founder & Chief Executive Officer (CEO). Ex-Google, YouTube & Hudson River Trading (HRT) Software Engineer. Training Neural Foundation Models. LinkedIn Profile · GitHub Profile
Marzieh Zare, PhD — Co-Founder & Chief Science Officer (CSO). Physics PhD in complex systems, neural criticality, and spontaneous brain network dynamics. LinkedIn Profile · Google Scholar Publications

Contact & Investment Channels

Seed-stage investor inquiries: invest@neurogenis.health

Scientific collaboration & data partnerships: research@neurogenis.health

Full technical dossier and scientific sources available at /about