辅导 ENG5059P Preliminary Report on MSc Project讲解 Python编程

Preliminary Report on MSc Project

Please start by saving this file with the name:

2961024_Jiayi Tang_ENG5059P_Prelim_2024-2025 Please tick appropriate project:

ENG5059P (MSc)

1. Introduction

Electroencephalography (EEG) detects electrical activity generated by synchronized neuronal firing in the brain. When neurons communicate, they create tiny electrical currents (10-100 microvolts) that travel through brain tissue, skull, and scalp. These signals are captured by scalp electrodes as oscillating voltage patterns.

EEG-Emotion Correlations

Research confirms distinct EEG signatures for basic emotions:

●     Calm State: Dominant posterior alpha waves (8-12 Hz) - rhythmic spindle-like waveforms indicating relaxed wakefulness

●     Acute Stress: Elevated frontal beta activity (13-30 Hz) - desynchronized fast oscillations signaling cognitive hyperarousal

●     Positive Emotion: Left-frontal alpha suppression (>15% power asymmetry) showing approach motivation

Educational Accessibility Gap

Despite EEG's potential, hands-on teaching faces critical barriers:

a) Complex setups (≥32 channels) deter 87% of engineering labs from including EEG practicals

b) Advanced processing requirements create steep learning curves for undergraduates

c) No standardized protocols exist for basic emotion detection

Approach

A simplified framework using:

a) Video-based emotion induction (YouTube videos)

b) 2-channel EEG (Fp1/Fp2 only)

c) Basic feature extraction (Alpha/Beta power)

d) Linear classification (LDA)

Educational Value

Provides replicable teaching materials for engineering education, enabling students to complete EEG emotion detection in one lab session.

2. Aims & Objectives

Primary Aim

To validate a 2-channel EEG setup for differentiating calm, stress, and happiness in student cohorts using frontal alpha/beta oscillations.

Objectives

A) Design video-based emotion induction protocol

●     Calm: Nature landscapes (5 min)

●     Stress: Exam countdown timer (5 min)

●     Happiness: Animal comedy clips (5 min)

B) Collect EEG data from 10 students

●     Channels: Fp1/Fp2 only

●     Parameters: 250Hz sampling, 1-30Hz bandpass

C) Extract spectral features

●    Alpha power (8-12Hz)

●     Beta power (13-30Hz)

D) Classify emotions with LDA

●     3-fold cross-validation

●     Target accuracy: >33% (chance level)

3. Resources Required

4. GANTT Chart

5. Risk Assessment

A)  Video-Induced Distress

Risk Level: Low

Control Measures:

All video content is previewed to exclude extreme material