flowRS Open the IDE
← Courses

Signals of the Brain

Open a real EEG, see the alpha rhythm appear when the eyes close, average a response to a cue and remove mains hum — each step checked against MNE-Python. For research, education and engineering development; not a medical device. Recordings: PhysioNet EEG Motor Movement/Imagery Dataset (ODC-By 1.0). Text adapted from Delta-Predictive Biosensing's learning corpus, © 2026 AuraSense Tech Corporation, CC BY 4.0.

Who this is for: Students and researchers starting with EEG; engineers who record from people. No neuroscience background needed.

01

Open an EEG

A recording is channels, a sample rate, units and events — read from a file that never leaves the browser.

Written
  1. Load the lesson "Open an EEG" in the IDE. It puts a minute of PhysioNet's eyes-open baseline (subject 1, six scalp channels at 160 Hz) in the browser's own file store; nothing is uploaded.
  2. Read the printout: the format (EDF+), the channels, the rate and the length. The same reader opens BDF, GDF, WFDB, XDF and BrainVision. On their test recordings EDF, BDF, WFDB, XDF and BrainVision match pyedflib, edfio, wfdb-python, pyxdf and MNE sample for sample.
  3. Open the Front panel to see each channel's name and unit (microvolts, as the file says), and the run's one annotation.
  4. Keep only O1, Oz and O2 by naming them in the node's channels.
  5. Import a recording of your own in the file panel and press Open recording. If its header names a person, the node warns, and for EDF and BDF files Anonymize writes a copy without the name.

Outcome: You can open a recording in flowRS and say what it holds.

02

Alpha rises with eyes closed

The 8–12 Hz alpha rhythm grows over the back of the head when the eyes close; measure it with Welch's method.

Written
  1. Load "Alpha rises with eyes closed". The eyes-open and eyes-closed baselines of the same volunteer each go through Band Power over O1, Oz and O2.
  2. Band Power is the Welch spectrum at MNE's defaults (256-point FFT, Hamming window, no overlap) summed over 8–12 Hz.
  3. Compare the two means: on this recording the eyes-closed alpha is about 16 times the eyes-open alpha, the figure MNE gives on the same data.
  4. Move both bands to 18–25 Hz and the difference shrinks: the large rise with the eyes closed is in the alpha band.

Outcome: You can measure band power and show the alpha rhythm in a real recording.

03

Average an event-related response

A response too small to see in one trial becomes plain in the average of many.

Written
  1. Load "Average an ERP". The recording is two minutes of the fist task: cue T1 asks for the left fist, T2 for the right.
  2. Epochs cuts 0.2 s before to 0.5 s after each T1 cue and removes each epoch's mean before the cue, as MNE's Epochs does at its defaults.
  3. Evoked Average averages the epochs and reports how many went in.
  4. Switch the events to T2 and compare the response over C3 with C4.

Outcome: You can cut epochs around events, baseline-correct them and average them.

04

Remove line noise

Mains wiring hums at 60 Hz (50 Hz in much of the world); a notch removes one narrow band and leaves the rest.

Written
  1. Load "Remove line noise". Band Power measures 59.5–60.5 Hz before and after a Notch.
  2. The Notch is MNE's default FIR notch: one zero-phase band-stop with a stop band per frequency, so nothing is delayed. Its method iir gives MNE's IIR notch.
  3. Change the notch's width and compare the line band before and after it.
  4. Put a Resample to 100 Hz before the Notch and the Notch refuses: Resample filtered out everything above the new 50 Hz Nyquist frequency, 60 Hz included, before it could fold back.

Outcome: You can remove mains interference without distorting the rest of the signal.

Open the IDE