A running log of what we're building, testing, and learning. Published as things happen — no polish, no press releases. The honest version.
Going public with progress. The goal: treat this like a wet-lab notebook anyone can read. Trainers, clinicians, and investors shouldn't have to wait for a pitch deck to see what we're working on — they should be able to watch the prototype take shape, week by week.
Next up: finishing the 2D pyEIT mesh, filing the provisional patent, and running the first Hilbert-Huang pass on a clean reference signal to baseline the denoising before we touch the phantom.
Got a 16-electrode 2D circular reconstruction mesh building end-to-end in pyEIT. Forward model solves in ~40ms on synthetic data. Inverse reconstruction still has edge artifacts near the boundary electrodes — expected, working on regularization next.
Ugly but functional. 16 electrodes on an adjustable strap, breadboarded multiplexer, bench power supply. Not going on anyone's arm yet — this is just to verify we can inject current across arbitrary pairs and measure the voltage differential everywhere else cleanly.
Early reading: SNR is noisier than I want at low injection amplitudes. Candidate fix list: better shielding on the measurement lines, moving to differential amps, and — the big one — letting HHT do its job on the output.
Deep dive into the original NASA HHT papers plus recent biomedical applications. The empirical mode decomposition step is the interesting one for us — unlike FFT it doesn't assume stationarity, which is perfect for the kind of quasi-random noise you get through muscle tissue.
Key takeaway: the intrinsic mode functions cluster predictably when the injected signal is clean. Noise ends up isolated in the highest-frequency IMFs and can be dropped without touching the signal band.
Opened conversations for the $60K pre-seed round. Structure: convertible notes, 20% discount, $4M cap. Target close by end of Q2. First meetings scheduled with two university-affiliated funds and one solo angel in the sports-tech space.
Three initial conversations with university sports-medicine departments about a potential pilot. Consistent theme: trainers want objective tissue data, not another wearable that spits out step counts. That's the wedge — replacing subjective palpation with a repeatable impedance map.
Sourcing comparison done. Tested three candidates for dry-contact electrodes: conductive silicone, stainless steel pads, and carbon-loaded fabric. Fabric wins on comfort + repeatability, silicone wins on conductivity. Leaning silicone for the phantom setup, fabric for the eventual wearable sleeve.
EIT-tek is live. Scope for the next 90 days: hardware v0, 2D reconstruction pipeline, HHT on example data, and a provisional patent filing. Everything after that — phantom validation, NASA licensing, pilot testing — waits on that foundation.