Milestone 3.2: Test

The group began testing by calibrating the OpenBCI hardware. This entailed fitting the headset and electrode modules in appropriate locations. Subsequently, the data aggregation and connection capabilities of the headset were confirmed. 

Following this, the group began focused testing on the promised capabilities of stress detection. The group used ChatGPT to generate short stories with keywords of "suspense" and "tragic" to assess the response of test subjects to emotional states evoked by the content. Particularly, the suspenseful story, as the tension gradually increased as the story progressed, provided valuable insight into the brain's response to emotionally stressful stimuli. These stories were subsequently tested on group members and other random samples of test subjects.

Putting aside the results of the prior testing, the group surmised that testing subjects' readings when watching movies, most notably the readings at certain climactic scenes of the movies, would produce relevant data. Some examples of movies the group used were Halloween, The Nightmare on Elm Street, and John Wick. Evidently, the sample size was much more limited as each movie is lengthy and we have limited hardware, however, the data gleaned was very discerning. Particularly, at longer duration emotional moments of tension, the brain was much more responsive than the short spikes at "jumpscare" type moments. 

With these rounds of testing, the group concluded that the electrode nodes that were centralized around the temporal lobe had the most significant reaction to emotionally stressful situations/stimuli. Another point of note is that the BCI gave a more comprehensive response to stressful situations rather than short bursts of fear from "jump scares". This gives our implementation of the OpenBCI a point of distinction with regular heart rate monitors and the like, as an increased heart rate is not always directly correlated to emotional stress. Our tests show that these distinct emotional events are uniquely discernible by our OpenBCI implementation and Psyche approach.

For our further development, we will focus on testing the brain's response to suspenseful moments as they are perfect microcosms to test response to emotional stress. In addition, the group will centralize the electrode nodes on the temporal lobe for a more detailed analysis of the most relevant division of the brain for Psyche's function.

Running OpenBCI API through Processing. Connecting to the headset through auto connect.

Through Processing, testing if user is relaxing or concentrating. Further work to be done to flesh out framework for more complex stress detection.

Milestone 3.3: Teamwork

Sankalp Patel: Customer Researcher

• Researched best individuals to test our system with and the specific functionalities needs to satisfy the needs of the users.

Pawan Perara: Industry Researcher

• Researched the divisions of the brain and the 10-20 system of EEG placement in order to discern the method of implementing the Psyche functionality using the OpenBCI.

Mitchell Reiff: Developer 

• Actively reverse engineering the open-source OpenBCI GUI widget code. Analyzing the OpenBCI documentation. 

Peter Shikhman: Industry Researcher

• Conducted research into professional methods of diagnosing stress. Used new knowledge to provide feedback for test cases.

Thomas Wang:  Research and Development

• Researching how other research groups have used the OpenBCI Ultracortex and finding ways to implement their findings into our own work.

Justin Young: Developer

• Working with Mitchell to analyze, test, and run OpenBCI GUI

Team Assessments