Power spectral analysis on resting state (eyes closed)dataset recorded over weeks

[oreHGA]

• [x] read raw brainwaves & signalquality values
• [x] split into epochs - 3-5seconds
• [x] average resting state for a single recording
• [x] then compare two datasets together. (this will be the basis for defining baseline vs. new dataset)
  • [x] see how data has changed over time - "morning", "afternoons", "evenings"

Dataset

• Resting state EEG data Link

• [x] upload dataset to hugging face

[oreHGA]

This is the basis for the next study we'd be running & sharing current data + analysis will hopefully ginger people to join in!

[oreHGA]

Erik has been working on a script that returns for a single recording :

• average powerByBand across channels
• average focus, calm predictions

Script

    return {
        "timestamp": df_pbb.index[0],
        "duration": df_pbb.index[-1] - df_pbb.index[0],
        "avg_power_per_channel_by_band": {
            channel: {band: df_pbb[channel + "_" + band].mean() for band in bands}
            for channel in channels
        },
        "avg_power_by_band": dict(average_band_power),
        "avg_power_by_channel": dict(average_channel_power),
        "avg_calm_score": avg_calm_score,
        "avg_focus_score": avg_focus_score,
        "time_spent_calm": time_spent_calm,
        "time_spent_focused": time_spent_focused,
        # `relative_power` keys are 2-tuples (band1, band2), values are ratios
        # maybe doesn't need to be computed here,
        # can be computed later from `avg_power_by_band`
        # "relative_power": {},
    }

Oh I think we could also include signal quality averages per channel

[oreHGA]

There's a snippet of code that allows us to choose channels based on signal Quality