Jazlin
Neuroengineer
Explainable Artificial Intelligence (XAI)
Goal: Apply GDAR (Graph Diffusion AutoRegressive) flow model to analyze NHP data, assessing trial-by-trial variability.
Responsibilities:
• Discern dominant GDAR flow patterns of slow and fast trials using statistical measures as an exploratory analysis.
• Implemented machine learning techniques like dimensionality reduction, spectral clustering, and feature engineering to extract underlying connectivity patterns from NHP data.
Smart Music
Goal: Train a closed-loop AI to generate music based on quantified enjoyment response information from the user’s EEG data.
Responsibilities:
• Built an EEG headset by 3D-printing the skeleton and tips for recording electrodes to connect to an OpenBCI Cyton board.
• Created a Python program that streams raw EEG data while performing real-time signal processing to omit physiologic artifacts and analyze power in each band.
• Interpret and quantify processed EEG data to train an ML algorithm that generates music based on user response.
ElectroMyoPathy
Goal: Develop an easy-to-use assistive communication device using biosignals for use in acute care settings between patient and care team.
Responsibilities:
• Designed a feedback system to collect raw accelerometer and EMG data through Arduino interface and communicate with React.
• Collaborated with clinical physicians and patient to test and optimize prototype at UWMC Harborview.
• Led my team to win the CNT Neural Engineering Tech Studio competition and was rewarded with resources to incubate project as a startup.
Neuropixel Analysis in the Primary Visual Cortex
Goal: Integrate rigorous quality metrics to spike sorting algorithms to understand the visual signal processing in V1 with a neuropixel probe.
Responsibilities:
• Utilized signal processing techniques on action potentials recorded from neuropixel probes.
• Applying advanced statistical analyses to assess the spike sorting algorithm's performance.
• Currently engaged in ongoing research with a focus on refining the spike sorting algorithm.
Explainable Artificial Intelligence (XAI)
Goal: Apply GDAR (Graph Diffusion AutoRegressive) flow model to analyze NHP data, assessing trial-by-trial variability.
Responsibilities:
• Discern dominant GDAR flow patterns of slow and fast trials using statistical measures as an exploratory analysis.
• Implemented machine learning techniques like dimensionality reduction, spectral clustering, and feature engineering to extract underlying connectivity patterns from NHP data.
Skills Learned:
Research Artefacts:
Adaptive Deep Brain Stimulation (DBS) in Obsessive-Compulsive Disorder (OCD)
Goal: Use 3D pose estimation in an acute hospital setting to analyze behavioral data during neuromodulation optimization.
Responsibilities:
• Generate a digital sync line through a microcontroller and serial communication to synchronize 6 cameras in real-time.
• Assembled and troubleshooted the custom PCB implementation with the camera hardware and software on Ubuntu.
Skills Learned:
Research Artefacts:
Neuropixel Analysis in the Primary Visual Cortex
Goal: Integrate rigorous quality metrics to spike sorting algorithms to understand the visual signal processing in V1 with a neuropixel probe.
Responsibilities:
• Utilized signal processing techniques on action potentials recorded from neuropixel probes.
• Applying advanced statistical analyses to assess the spike sorting algorithm's performance.
• Currently engaged in ongoing research with a focus on refining the spike sorting algorithm.
Skills Learned:
Research Artefacts:
Smart Music
Goal: Train a closed-loop AI to generate music based on quantified enjoyment response information from the user’s EEG data.
Responsibilities:
• Built an EEG headset by 3D-printing the skeleton and tips for recording electrodes to connect to an OpenBCI Cyton board.
• Created a Python program that streams raw EEG data while performing real-time signal processing to omit physiologic artifacts and analyze power in each band.
• Interpret and quantify processed EEG data to train an ML algorithm that generates music based on user response.
Skills Learned:
Research Artefacts:
ElectroMyoPathy
Goal: Develop an easy-to-use assistive communication device using biosignals for use in acute care settings between patient and care team.
Responsibilities:
• Designed a feedback system to collect raw accelerometer and EMG data through Arduino interface and communicate with React.
• Collaborated with clinical physicians and patient to test and optimize prototype at UWMC Harborview.
• Led my team to win the CNT Neural Engineering Tech Studio competition and was rewarded with resources to incubate project as a startup.
Skills Learned:
Research Artefacts: