πŸ“— -> EEG Witness Testimony Assessment


We now have a Notion Page for readings, put there as well.

Pages

EEG-Oscillations-Reflect-Cognitive-and-Memory-Performance
Review-Retention-Recall-Processes
Analysis-of-Eyewitness-Testimony-Using-Electroencephalogram-Signals

Future Reading

🎀 Vocab

πŸ’‘-> Ideas

Questions:

This paper has a much higher sampling rate (2048 Hz). Even after it is downsampled to 512 Hz, it is still higher than the sampling rate of the CytonBoard (only 250 Hz). This is something to keep in mind when applying filtering techniques (know that your data will look different and thus must be filtered/treated differently).

  • How would this affect our experiment?

Current Pipeline

Transclude of Witness-Reading-2024-12-06-01.54.13.excalidraw

Options

Transclude of Witness-Reading-2025-04-15-11.17.00.excalidraw

βœ’οΈ -> Reading Notes

Analysis of eyewitness testimony using electroencephalogram signals

Design

Used P. Bem-Haja’s dataset of 20 second crime videos, and asked participants question

Analysis

Selected theta, alpha, and beta bands.

  • Theta waves - associated with mental operations: attention, learning, memory recalls
  • Beta waves - active, busy, anxious thinking
  • Alpha waves -

Finally, studies like 6 have been made concerning the asymmetry of the alpha waves in questions of emotion and motivation.

Gamma Oscillations Distinguish True From False Memories

Encoding

Increased gamma power (28–100 Hz) in the hippocampus, prefrontal cortex, and left temporal lobe predicted whether an item would later be successfully recalled.

Retrieval

A similar increase in gamma power during the 500 ms preceding correct recall was observed in the same regions. False recalls exhibited a weaker or absent gamma-power response in these regions.

Effect of EEG evidence on juror decision

Takeaway:

  • Not directly useful for the project, but an interesting ethical consideration
    • The evidence of EEG lie detection swung a decision by 20%, 40% β†’ 20%
  • Is it ethical to use/present evidence with a degree of uncertainty in trial

Resources

Connections