Cortical recording lab guide

Controls, experiments, and interpretation.

Open the lab →

Compare recording geometries

  1. Choose Single contact, Depth probe, or 2 × 2 array.
  2. Rotate with left-drag, pan with right-drag, and zoom with the wheel. Top/Side view buttons provide repeatable viewpoints.
  3. Select an electrode with Prev/Next. Set its coordinates or click the slab; Shift-click also sets depth.
  4. Vary noise drive, pause, and inspect the two-second raster and traces. Moving a contact clears old recordings.
  5. Export raw CSV to save every contact's unfiltered samples at 8 kHz. Restart recording repeats the seeded model with the current parameters and contacts.

Interpretation

The raster uses known synthetic source events, selected by an expected isolated peak of at least 6 µV. It is not a detector applied to the noisy trace. The trace reference line is display-only. RMS describes the selected unfiltered two-second buffer; initially it includes empty prehistory.

Model

There are 1,600 seeded point neurons, stochastic spike events, local recruitment, distance-weighted spike waveforms, and correlated background noise. Fixed 4 ms time steps carry waveform tails into the next block. Contact distance changes amplitude without adding a transit delay; the template trough follows its source event by 0.55 ms. Source firing and contact noise use separate seeded random streams, so passive contact selection and noise settings do not change the neurons’ events. Noise follows a contact ID when supplied, or its coordinate identity in this editor. Signal units are synthetic µV and are not calibrated to biological recordings. The model omits layered conductivity, membrane biophysics, impedance, and a physical reference electrode.