a 600 channel wearable multimodal fNIRS-EEG headband with real-time attention decoding

System description: Neurode is a head-worn, self-contained combined EEG-fNIRS- headband that records over 600 channels of hemodynamic and electrophysiological activity across the entire cortex. Unlike benchtop or cart-based optical systems, the entire optical front end, acquisition electronics, and streaming stack live in a single headband that fits like a normal piece of consumer hardware and can be placed on a new user in roughly 30 seconds. Data streams continuously to a laptop, where it is processed, visualized, and decoded in real time.

What the demo involves: Attendees will be fitted with the headband and see their own brain activity in real time.

The demo has two layers: 1. Live 600-channel visualization. Oxy- and deoxy-hemoglobin changes and electrophysiological activity projected onto an anatomical map of the brain, updating continuously. Attendees can watch their own signal respond to simple manipulations — a breath hold, mental arithmetic, closing their eyes etc. 2. Locus-of-attention inference. A decoder that classifies, moment to moment, whether the wearer's attention is directed inward or outward. We guide attendees through a few short tasks that alternate between the two — reading a passage or scanning the room for external attention, mental arithmetic or recalling a memory with eyes closed for internal — and the decoder's output is displayed live alongside the brain map, so attendees can see the classification shift as they switched.

Each run of the visualisation takes about two minutes end to end, and the attention coder takes another 5 minutes, so we can cycle a large number of attendees through the visualisation session or the decoder, depending on demand.

Why it's relevant to this workshop: Most high-density optical neuroimaging is still constrained to a lab. A wearable form factor at this channel count changes what kind of data can be collected and at what scale, which is directly upstream of the data problem that AI for neural data runs into. Internal versus external attention is also a useful decoding target: it's behaviorally meaningful, hard to fake, and not well served by existing consumer neurotech. We're the only portable EEG-fNIRS headband with this channel density, at this level of form factor, its portable, lightweight, and comfortbale to wear for long periods while maintaining a dense frontal montage.