EEG Headset

By shrijeetverma13 · March 23, 2026

Data source:

There are 344 extreme outlier readings in the P channel, identified using the 3×IQR method with a threshold of 4,671.31. These outliers range from 4,671.79 up to the maximum value of 362,564. A bar chart has been generated showing how these outliers are distributed across eye states (open vs. closed).

The analysis successfully compared EEG signal variance across 14 channels for both eye open and eye closed states. A grouped bar chart was generated showing the variance for each channel side by side, making it easy to visually identify which channels exhibit the most signal variability during each eye state.

The EEG channel with the largest mean difference between eye open and eye closed states is AF4 , with a mean absolute difference of 89.42 . This suggests that the AF4 channel (located in the right frontal area of the brain) is most sensitive to changes in eye state in this dataset.

The grouped bar chart and data table show how average EEG signal values differ across all 14 channels between eyes-open (State 1, 8,257 samples) and eyes-closed (State 2, 6,723 samples). Overall, the differences are relatively small compared to the absolute signal values (which range roughly from 4,000 to 4,700), but some channels do show notable variation.

Frontal channels (AF3, AF4, F3, F4) better distinguish eye states compared to occipital channels (O1, O2). The bar chart shows Cohen's d effect sizes for each channel, where higher values indicate stronger discrimination between open and closed eye states.

Both temporal channels T7 and T8 were analyzed across eye open and closed states. T7 has a higher baseline signal magnitude (~4,341 µV) compared to T8 (~4,230–4,233 µV). However, the difference between eye open and closed is minimal for both channels. T8 shows a slightly larger shift of 3.61 µV between states, while T7 barely changes (-0.03 µV). Two data tables were generated summarizing these comparisons.

The analysis reveals clear differences in signal stability across EEG channels. O2 has the narrowest signal range (2,696.92 µV) making it the most stable channel, while AF4 has the widest range (714,530.85 µV) indicating extreme volatility — though this is likely influenced by artifact spikes rather than true brain activity. A data table was generated showing the full ranking of all 14 channels from narrowest to widest range.

The analysis examined hemispheric asymmetry between left-right electrode pairs (AF3 vs AF4, F3 vs F4, F7 vs F8) across different eye states. Two data tables were generated summarizing the asymmetry metrics and statistical comparisons for each electrode pair.

Extremely few F8 readings fall below 1000 — just 2 out of 14,980 total readings, which is only 0.01%. Both of these low readings are exclusively associated with eye state 1 (100% of the below-1000 group), suggesting a potential link, though the sample is too small to draw strong conclusions. A bar chart has been generated showing the eye state distribution for these readings.

The median EEG signal profile across all 14 channels is remarkably similar between the two eye states. A grouped bar chart and data table have been generated showing the median signal values for each channel. For Eye State 1, values range from 4,007.18 µV (F7, lowest) to 4,619.49 µV (P, highest). For Eye State 2, values range from 4,003.59 µV (F7, lowest) to 4,615.90 µV (P, highest). The differences between the two eye states are very small across all channels, with the largest difference occurring at the AF4 channel (only 7.69 µV apart).

The analysis examined correlations between all 14 EEG channels during the eye-closed state (eye state=2), using 6,723 samples. A heatmap and a bar chart were generated showing the strength of correlations across all channel pairs, along with detailed data tables ranking the top 20 most strongly correlated pairs.

A data table was generated showing the skewness of signal distributions across all 14 EEG channels, broken down by eye state. The table lets you compare how asymmetric each channel's signal distribution is overall and whether that asymmetry changes depending on the eye state recorded.

FC5 and FC6 display a notably asymmetric relationship that changes depending on eye state. When eyes are open (State 1), the two channels are nearly identical — FC5 averages 4200.39 and FC6 averages 4200.26, with a negligible asymmetry of just -0.13. However, when eyes are closed (State 2), FC6 rises to 4205.15 while FC5 drops to 4121.41, creating a much larger asymmetry of +83.74. This suggests the frontocentral region exhibits state-dependent lateralization rather than stable symmetry. Data tables summarizing these comparisons were generated for further review.

The analysis successfully identified the most strongly correlated EEG channel pairs during the eye-open state. A correlation heatmap and a bar chart of the top 20 channel pairs were generated, along with detailed data tables showing pairwise correlations across all 14 EEG channels (AF3, F7, F3, FC5, T7, P, O1, O2, P8, T8, FC6, F4, F8, AF4).

The analysis ranked all 14 EEG channels by their Cohen's d effect size to determine which best predict eye state. A bar chart was generated showing all channels ranked from highest to lowest effect size, with the top 5 highlighted in red. The visualization clearly shows which channels have the strongest separation between the two eye states, making them the best predictors.

After removing outliers from the P and AF4 channels using the IQR method, 12,578 of 14,980 rows (84%) were retained. The mean EEG channel values are very similar between eye open and eye closed states, but there are consistent small differences. Two bar charts and data tables were generated showing these comparisons clearly.