Content and Conceptual Learning Goals
Students will:
1. Master ERP methodology: Understand the complete pipeline from raw EEG data to publication-ready ERP waveforms
2. Apply electrophysiological theory: Connect specific ERP components to cognitive processes
3. Integrate literature: Synthesize research findings about their chosen ERP component with their own analytical results
4. Understand signal processing: Apply filtering, artifact rejection, averaging, and baseline correction with theoretical justification

MATLAB Learning Goals and Its Role
Students will:
1. Develop data analysis workflows: Create a complete, reproducible analysis pipeline in MATLAB Live Script format
2. Master scientific programming: Write clear, well-commented, efficient code following best practices
3. Use specialized toolboxes: Apply EEGLAB functions for neurophysiological data analysis
4. Create quality visualizations: Generate professional ERP waveforms, topographic maps, and statistical plots
5. Debug and troubleshoot: Identify and resolve coding errors independently

Why MATLAB improves student learning:
- Live Script format: Integrates narrative explanation, code, and output, promoting computational literacy
- EEGLAB integration: EEGLAB is a standard toolbox used in ERP/EEG research
- Reproducibility: Students create shareable, executable documents that demonstrate their complete analytical process
- Professional relevance: MATLAB is widely used in cognitive neuroscience research labs
- Immediate feedback: Students can visualize results of analytical decisions in real-time, deepening conceptual understanding

Could this be done with other software? Yes, Python (with MNE-Python) or R could be used, but MATLAB with EEGLAB was chosen because:
- EEGLAB is the most widely-used EEG analysis package in cognitive neuroscience
- Live Script format uniquely combines documentation and code
- MATLAB's visualization capabilities are clear and versatile
- Most cognitive neuroscience labs students will enter in the future will use MATLAB/EEGLAB

Higher-Order Thinking Skills
Students develop:
1. Critical thinking: Evaluate and justify analytical decisions; assess methodological limitations
2. Computational thinking: Decompose complex analyses into logical steps; recognize patterns in data processing
3. Scientific reasoning: Form hypotheses based on theory; interpret results in light of predictions
4. Problem-solving: Troubleshoot code errors; adapt methods when initial approaches fail
5. Synthesis: Integrate theoretical knowledge, methodological understanding, and practical skills
6. Metacognition: Reflect on learning process and use of AI assistance (self-reflection essay)

Additional Skills Development
1. Scientific writing: Compose clear theoretical background and results interpretation
2. Oral presentation: Communicate complex methods and findings to peers (15-minute presentation with Q&A)
3. Peer review: Provide constructive feedback on draft Live Scripts
4. Information literacy: Find and evaluate primary research literature
5. Data visualization: Design figures that effectively communicate results
6. Project management: Plan and execute a multi-week independent project with milestones
7. Academic integrity: Use AI tools responsibly and acknowledge assistance appropriately

Educational Setting
This activity is designed for an upper-level undergraduate seminar course (PSYCH 390: Electrophysiology of Complex Cognition) at a liberal arts college (St. Olaf College). The course integrates critical analysis of electrophysiological research with hands-on methodological training, examining how EEG and ERPs reveal neural mechanisms underlying attention, memory, language, and decision-making.

Course Prerequisites
- PSYCH 225, PSYCH 237 or 238, or consent of instructor
- Students should have foundational knowledge of cognitive psychology and research methods
Class Structure
- Typical class size: 12-20 students (seminar format)
- Activity type: Final project spanning 5-6 weeks
- Time commitment:
o In-class workshops: ~6 hours across semester
o Independent work: 30-40 hours
o Peer review: 1-2 hours
o Presentation: 15 minutes per student
Required MATLAB Skills (Students Must Have Before Starting)
Students must be comfortable with MATLAB before beginning this project, specifically:
- Basic MATLAB syntax and variable operations
- Writing scripts and functions with comments
- Creating basic plots and figures
- Understanding data structures (structs, arrays)
- Using MATLAB Live Script format
- Installing and loading toolboxes (EEGLAB)
- Debugging code and interpreting error messages
These skills are developed through instructor-led tutorials and collaborative workshops during weeks 1-9 of the semester.
Required Disciplinary Knowledge
Before starting the project, students should understand:
- Basic principles of electrophysiology and neural signaling
- Cognitive processes (attention, memory, language, decision-making)
- Experimental design in cognitive psychology
- Signal processing concepts (filtering, noise, artifacts)
- Statistical analysis fundamentals
- Scientific writing and literature review skills
Course Integration
This final project synthesizes skills developed throughout the semester:
- Weeks 1-3: Introduction to EEG/ERP methods and cognitive neuroscience theory
- Weeks 4-6: MATLAB basics and EEGLAB tutorials
- Weeks 7-9: Preprocessing techniques and analysis pipelines
- Weeks 10-15: Independent project work with scaffolded milestones
Adaptability to Other Settings
Difficulty: Moderate to High
This activity is specifically designed to integrate with the cumulative learning of a semester-long electrophysiology course, making direct transfer challenging. However, it could be adapted for:
- Other neuroscience/psychology courses: With modification to focus on different neuroimaging methods (fMRI, MEG)
- Biomedical signal processing courses: By emphasizing signal processing over cognitive theory
- Data science courses: By treating ERP data as a signal processing/classification problem
- Graduate seminars: By increasing expectations for methodological sophistication and literature review depth
Key requirements for adaptation:
- Access to EEG/ERP datasets (publicly available datasets can substitute)
- MATLAB licenses with EEGLAB toolbox
- Instructor expertise in electrophysiology or willingness to learn alongside students
- Sufficient course time for MATLAB skill development (or students with existing MATLAB experience)

Overview of the Activity
This final project requires students to complete a full ERP analysis from raw data to interpretation, documenting their process in a MATLAB Live Script. The project emphasizes not just completing the analysis, but understanding and justifying every analytical decision. Students present their findings orally to demonstrate genuine comprehension, even if they used AI coding assistance.
How MATLAB is Used
MATLAB serves as the primary tool for:
1. Loading and inspecting EEG data - Students use EEGLAB functions to import provided datasets and examine data structure
2. Preprocessing pipeline - Implementing filtering, artifact rejection, re-referencing, epoching, and baseline correction
3. ERP generation - Averaging trials and calculating standard errors
4. Visualization - Creating ERP waveforms and topographic scalp maps
5. Statistical analysis - Performing comparisons between conditions
6. Documentation - Live Script format integrates code, output, and narrative explanation
Live Script integration is crucial - students write theoretical explanations, show code with extensive comments, display output figures, and provide computational justifications all in a single, executable document. This format uniquely supports the learning goal of connecting theory, computation, and interpretation.
Project Timeline and Structure
Week 10: Component Selection
- Students submit 1-paragraph proposal (200-250 words) choosing one ERP component (P300, N400, N170, ERN, MMN, N2pc, LRP, or P600)
- Proposal must justify choice based on cognitive interest and research relevance
- Instructor provides feedback within one week
Weeks 11-13: Independent Work
- Weekly lab sessions for troubleshooting and skill development:
o Week 11: Data loading and inspection workshop
o Week 12: Preprocessing troubleshooting clinic
o Week 13: Creating publication-quality figures
- Students work through analysis pipeline independently
- Office hours available for individual consultation
Week 14: Draft Submission and Peer Review
- Complete draft Live Scripts due (5% of grade)
- In-class peer review workshop
- Students exchange feedback on structure, code quality, and clarity
- Instructor provides formative feedback
Week 15: Revision Period
- Students incorporate feedback
- Practice presentations can be scheduled
- Individual consultations available
Finals Week: Final Deliverables
- Final Live Script submitted (65% of grade)
- 12-minute presentations + 3 minutes Q&A (15% of grade)
- Self-reflection essay submitted (5% of grade)
Required Live Script Components
The final MATLAB Live Script must include six main sections:
Section A: Introduction and Theoretical Background (15%)
- 750-1000 words with minimum 5 peer-reviewed sources
- Characterize the ERP component (latency, polarity, scalp distribution)
- Explain neural generators and cognitive function
- Review 3-4 key research studies
- Describe classic experimental paradigm
- State predictions for the provided data
Section B: Data Loading and Initial Inspection (10%) Students write MATLAB code to:
- Load provided .set or .mat EEG dataset
- Display dataset properties (sampling rate, channels, trials)
- Show channel locations on topographic map
- List event codes and verify data structure
- Plot raw continuous EEG from multiple channels
- Identify potential artifacts visually
Critical element: For each code block, students must explain WHY this inspection step matters and what it reveals about data quality.
Section C: Comprehensive Preprocessing Pipeline (40%) This is the core of the project. Students implement and justify:
Step 1: Filtering (8%)
- Implement high-pass and low-pass filters
- Justify cutoff frequencies for specific component (e.g., "P300 requires 0.1-30 Hz because...")
- Specify filter type and order
- Show power spectral density before/after
- Discuss potential filter artifacts
- CRITICAL THINKING: What information is lost? What artifacts introduced?
Step 2: Channel Operations (6%)
- Detect bad channels algorithmically and visually
- Interpolate bad channels using spherical spline
- Re-reference data (justify reference choice for component)
- Visualize effects of re-referencing
- JUSTIFICATION: Why this reference? Why interpolate vs. exclude?
Step 3: Artifact Detection and Rejection (8%)
- Set voltage thresholds for artifact detection
- Implement gradient criteria for muscle artifacts
- Show examples of rejected trials
- Display rejection statistics across conditions
- CRITICAL ASSESSMENT: Are thresholds appropriate? How does rejection affect statistical power?
Step 4: Epoching (6%)
- Extract event-related time windows
- Justify epoch length for component (must include baseline + full component)
- Show epoch counts per condition
- CONSIDERATION: How does window affect frequency content?
Step 5: Baseline Correction (6%)
- Apply baseline correction (typically -200 to 0 ms)
- Justify baseline period choice
- Visualize effects on waveforms
- CRITICAL THINKING: What assumptions? When problematic?
Step 6: Averaging (6%)
- Average trials within conditions
- Calculate standard error/confidence intervals
- Display final trial counts after rejection
- JUSTIFICATION: Why does averaging improve SNR? What's assumed about variability?
Section D: ERP Visualization and Measurement (20%)
Publication-Quality Figures Required:
1. ERP waveforms - Grand averages with multiple conditions overlaid, time on x-axis, amplitude (µV) on y-axis, shaded confidence intervals, clear legend, component window marked
2. Topographic maps - Scalp distribution at peak latency, appropriate color scale, multiple time points/conditions
3. Component measurements - Bar plots comparing conditions with error bars
Measurement approach:
- Define strategy (peak amplitude, mean amplitude, or peak latency)
- JUSTIFY: Which measure is most appropriate for this component and why?
- Extract values for statistical analysis
Section E: Statistical Analysis (10%)
- Choose appropriate test (t-test, ANOVA based on design)
- Justify test choice
- Report results with effect sizes (Cohen's d, eta-squared)
- Visualize statistical comparison
- Interpret: Do results support predictions?
Section F: Results Interpretation and Critical Discussion (5%)
- 500-750 words
- Summarize findings in relation to theory
- Compare to previous literature
- Address methodological limitations
- Suggest future improvements
- Discuss broader implications for understanding cognition
Assessment Components
1. Draft Live Script (Week 14) - 5%
- Graded on completeness and effort
- Must include all sections and runnable code
- Provides opportunity for feedback before final grading
2. Final Live Script (Finals Week) - 65% Detailed rubric evaluates:
- Theoretical understanding (15%)
- Data loading & inspection (10%)
- Preprocessing pipeline (40%)
- ERP visualization (20%)
- Statistical analysis (10%)
- Discussion (5%)
3. Oral Presentation (Finals Week) - 15% 12 minutes + 3 minutes Q&A
Required slides:
1. Title and component introduction 2-3. Theoretical background (what and why) 4-6. Methods walkthrough (emphasize 2-3 key decisions with justifications, show example code) 7-9. Results (ERP plot, topography, statistics)
2. Interpretation and implications
3. Limitations and future directions
Assessment focuses on:
- Can student explain decisions without reading slides?
- Does student understand what their code does?
- Can student answer questions about parameters?
- Can student defend methodological choices?
- Does student connect findings to theory?
Purpose of presentation: Ensures genuine understanding even if AI tools assisted with coding. Students must independently explain and justify every decision.
4. Self-Reflection Essay (Finals Week) - 5% 300-400 words addressing:
- Most challenging aspect
- Surprising lessons about ERP analysis
- How AI assistance was verified (if used)
- How to approach similar analyses differently in future
Graded on thoughtfulness and honesty, not specific answers.
Academic Integrity and AI Usage
Explicitly Permitted:
- Using AI tools (ChatGPT, MATLAB Copilot) for debugging
- Consulting documentation and online tutorials
- Discussing approaches with classmates
- Receiving feedback on code structure
Required:
- Must understand every line of code
- Must cite substantial borrowed code
- Must acknowledge AI assistance in reflection
- Must be able to explain in presentation
Prohibited:
- Submitting code without understanding
- Copying complete pipelines without modification
- AI-writing theoretical sections
- Misrepresenting extent of assistance
Key point: The oral presentation serves as an oral exam. Students who cannot explain their work will have grades reduced significantly regardless of Live Script quality.
Materials Provided to Students
1. EEG Dataset
- EEGLAB-compatible format (.set/.fdt or .mat files)
- Pre-processed to remove only catastrophic artifacts
- Multiple experimental conditions for comparison
- Documented event codes
- Channel location information included
- Sufficient trials per condition for robust ERPs
2. Template Live Script
- Structural outline with section headers
- Placeholder text for required content
- Example code comments showing expected documentation level
- Does NOT include complete analysis code (students must write this)
3. Example Analysis Script
- Complete working example using a DIFFERENT component than students will analyze
- Demonstrates expected level of commenting and justification
- Shows proper visualization formatting
- Illustrates Live Script integration of narrative and code
4. Research Literature
- 3-5 seminal papers for each component option
- Papers demonstrate classic experimental paradigms
- Include both original descriptions and recent applications
- Accessible through library databases
5. Technical Documentation
- EEGLAB installation instructions
- Quick reference guide for key EEGLAB functions
- MATLAB Live Script tutorial
- Visualization best practices guide (formatting, color choices, axis labels)
6. Tutorial Videos
- Data loading and structure examination (15 min)
- Filtering and preprocessing basics (20 min)
- Creating topographic maps (10 min)
- Exporting publication-quality figures (12 min)
7. Assessment Materials
- Complete rubric for Live Script (shared Week 10)
- Presentation rubric with example slides
- Peer review worksheet for draft feedback
- Self-reflection essay prompts
Support Structure
Weekly Lab Hours:
- Structured sessions with specific focus each week
- Peer collaboration encouraged
- Instructor available for troubleshooting
Online Discussion Forum:
- Students post questions and help each other
- Instructor monitors and provides guidance
- Code snippets can be shared (with proper attribution)
Individual Consultations:
- Available by appointment throughout project
- Especially encouraged before final submission
- Can practice presentations for feedback
Why This Activity Works
Integration of theory and practice: Students cannot complete the analysis without understanding the underlying electrophysiology and cognitive theory.
Emphasis on justification: Every analytical decision requires theoretical and computational justification, preventing "black box" application of methods.
Authentic assessment: The oral presentation reveals whether students genuinely understand their work, addressing concerns about AI-generated code.
Scaffolded independence: Milestones (proposal, draft, final) provide structure while allowing creative choice of component and interpretation.
Professional preparation: Mirrors authentic research workflow; creates portfolio piece for graduate school applications.
Computational literacy: Live Script format develops ability to communicate technical work clearly, an essential skill for modern scientists.

Tips for Successful Implementation
Before the Project Begins:
1. Build MATLAB confidence early - Students need weeks 1-9 to develop comfort with MATLAB basics before tackling this complex project. Don't rush the foundational tutorials.
2. Model the process - Show your own Live Script analyzing a different component during weeks 7-9. Walk through your reasoning for preprocessing choices.
3. Discuss data quality - Help students understand that real EEG data is messy. Not all analyses will produce textbook-perfect results, and that's okay. The justification matters more than beautiful waveforms.
4. Set realistic expectations - Students may need 30-40 hours outside class. Be explicit about this time commitment early.
Common Areas of Confusion
1. Filter Parameter Selection
- Confusion: "How do I know what cutoff frequencies to use?"
- Solution: Provide component-specific guidance. For example: "P300 literature typically uses 0.1-30 Hz, but you need to justify why based on the component's time course and frequency content."
2. Artifact Rejection Thresholds
- Confusion: "I'm rejecting too many/too few trials. How do I know if my thresholds are right?"
- Solution: Teach them to examine rejected trials visually. If clear artifacts remain, thresholds are too lenient. If clean data is rejected, too strict.
- Teaching tip: Typical starting points: ±75-100 µV for voltage threshold, but this varies by system and component. Students should show examples of accepted and rejected trials.
3. Re-referencing Choices
- Confusion: "Why does re-referencing change my waveforms so much?"
- Solution: Emphasize that there's no "true" reference in EEG - all measurements are relative. Different references highlight different aspects of the signal.
- Teaching tip: Have students plot the same data with 2-3 different references to see how interpretation might differ.
4. Baseline Correction Period
- Confusion: "Can I use any pre-stimulus period as baseline?"
- Solution: Baseline should be neutral - no systematic difference between conditions. Typically -200 to 0 ms, but students should verify conditions don't differ during baseline.
- Teaching tip: Have students plot baseline periods separately for each condition to check for pre-existing differences.
5. Peak vs. Mean Amplitude
- Confusion: "Which measurement should I use?"
- Solution: Peak is more influenced by noise; mean amplitude is more robust but requires defining a time window. Choice depends on component properties and literature norms.
- Teaching tip: Have students try both and compare - this develops critical thinking about measurement choices.
Things That Need Reinforcement
1. Justification, Not Just Execution Students often complete steps without explaining WHY. Repeatedly emphasize: "Your code works, but WHY did you choose these parameters?". This is the strength of a live script where they have the tools to describe WHY something works.
2. Code Comments Are for Humans Students may write minimal comments like "% filter data". Teach them to write: "% Apply 0.1 Hz high-pass to remove slow drifts while preserving P300's slow positive wave"
3. Connection to Theory Students sometimes treat this as a pure coding exercise. Continuously connect back to: "What does this preprocessing choice mean for your ability to measure the cognitive process you care about?"
4. Critical Thinking About Results If results don't match predictions, that's scientifically interesting! Teach students to thoughtfully discuss why results might differ from literature rather than assuming they "did it wrong."
Making the Best Use of the Activity
Leverage Peer Learning:
- The Week 14 peer review is crucial. Provide a structured worksheet asking reviewers to identify strengths, check that justifications are present, and suggest one area for improvement.
- Consider having students present work-in-progress during Week 12 or 13 for informal feedback.
Use Presentations Strategically:
- Schedule presentations over 2-3 days to prevent burnout
- Encourage students to ask each other questions - this builds classroom community and critical thinking
- Take notes on common strong points and struggles to inform future course iterations
Address AI Use Proactively:
- Don't treat AI as cheating - it's a tool. But emphasize verification and understanding.
- In the self-reflection essay, students who honestly engage with how they used AI and learned from it demonstrate more learning than those who claim not to have used it at all.
- The presentation is where AI use becomes obvious if students don't understand their code. Use this as a learning moment, not a punitive one.
Create a Supportive Environment:
- Share your own mistakes or struggles with EEG analysis
- Celebrate interesting "failures" - unexpected results that led to insights
- Display examples of excellent justifications from previous students (with permission)
Troubleshooting Technical Issues
Installation Problems:
- Have backup plan for students with MATLAB installation issues: MATLAB Online can run EEGLAB
- Create installation testing assignment in Week 1 to identify problems early
- Keep detailed troubleshooting documentation for common OS-specific issues
Data Loading Issues:
- EEGLAB has specific requirements for file paths and formats - provide clear documentation
- Have students verify data loaded correctly by displaying key properties
- Keep backup datasets in case files become corrupted
Memory/Performance Issues:
- EEG files can be large. Teach students to clear unnecessary variables and close figures
- If using student laptops, test that datasets are appropriately sized for available RAM
- Consider providing subset of full dataset if computational resources are limited
Code That Won't Run:
- Build in "code checkpoints" where students verify their analysis up to that point works
- Teach systematic debugging: read error messages carefully, check variable dimensions, isolate problem code
- Encourage use of MATLAB's built-in debugging tools (breakpoints, variable inspector)
Adapting for Different Student Backgrounds
For students with strong coding backgrounds:
- Encourage exploration of advanced visualization techniques
- Challenge them to implement additional analyses (e.g., time-frequency analysis)
- Have them mentor peers during lab sessions
For students struggling with MATLAB:
- Provide more extensive code templates (but still requiring justifications)
- Allow more time in office hours
- Consider allowing partnerships where students collaborate but submit individual Live Scripts with different components
For students with limited statistics backgrounds:
- Provide clearer guidance on which statistical test to use
- Focus assessment more heavily on correct implementation than on sophisticated test choice
- Supplementary statistics review materials
Assessment Tips
For Draft Review (Week 14):
- Focus on structure and completeness, not perfection
- Provide concrete, actionable feedback: "Your filtering justification needs more detail - explain how these frequencies relate to P300's time course"
- Identify students who are significantly behind and offer intervention
For Final Live Script:
- Use rubric consistently - resist grade inflation for "effort"
- Look for thoughtful justification even if analytical choices differ from your own
- Value critical thinking about limitations as much as perfect execution
For Presentations:
- Prepare 2-3 questions for each student in advance to probe understanding
- Be supportive but probe genuinely - "Can you explain why you chose that baseline period?" not "Is that the right baseline period?"
- Consider having students evaluate each other using a simplified rubric (teaches them to think critically about methods)
For Self-Reflection:
- Give full credit for genuine engagement with prompts
- Look for metacognitive insights, not self-deprecation or false confidence
- Use reflections to identify teaching improvements for next year
Timeline Warnings
Potential Bottlenecks:
- Week 14 draft submission can overwhelm if you have large class - stagger due dates or limit feedback depth
- Presentations in Finals Week compete with other final exams - consider end of Week 15 instead
- Student procrastination is common - consider required "check-in" meetings at Week 12
Long-Term Considerations
For Course Iteration:
- Save excellent examples (with permission) to show future students
- Document common questions to improve documentation/tutorials
- Track which components students choose - you may need to acquire more papers for popular ones
- Note which preprocessing steps cause the most confusion for targeted teaching next year
For Student Careers:
- This Live Script becomes a portfolio piece for graduate school applications
- Encourage students to present at undergraduate research conferences
- Strong projects could be developed into independent studies or honors theses
- Skills directly transfer to research assistant positions in neuroscience labs

How Student Learning Is Evaluated
Learning is assessed through multiple methods to ensure comprehensive evaluation of both understanding and execution:
1. Draft Live Script (5% - Week 14)
Purpose: Formative assessment providing feedback before final grading
Evaluation Criteria:
- Completeness: All sections present
- Code functionality: Script runs without errors
- Effort: Evidence of genuine engagement with each component
- Structure: Follows required section organization
Grading: Essentially completion-based. Students receive full credit for submitting a complete, working draft and participate constructively in peer review.
2. Final MATLAB Live Script (65% - Finals Week)
Purpose: Summative assessment of technical skills, theoretical understanding, and scientific communication
Detailed Evaluation Using Rubric:
Theoretical Understanding (15%)
- Quality and depth of literature review
- Accuracy of neural generator descriptions
- Sophistication of cognitive explanations
- Integration of multiple sources
- Clarity of predictions
Data Loading & Inspection (10%)
- Correct data import
- Thorough examination of data properties
- Insightful observations about data quality
- Appropriate visualizations
- Thoughtful justifications for inspection steps
Preprocessing Pipeline (40%) - Most heavily weighted
- Filtering: Appropriate parameters with sophisticated justification (8%)
- Channel operations: Correct implementation with rationale (6%)
- Artifact rejection: Thoughtful thresholds with quality assessment (8%)
- Epoching: Appropriate windows with theoretical justification (6%)
- Baseline correction: Correct implementation with critical thinking (6%)
- Averaging: Understanding of signal-to-noise improvement (6%)
For each preprocessing step, students must demonstrate:
- Theoretical rationale (why needed)
- Computational understanding (how it works)
- Parameter justification (why these values)
- Quality verification (does it work)
ERP Visualization (20%)
- Publication-quality waveforms with proper formatting
- Appropriate topographic maps
- Correct component measurement approach
- Professional figure presentation
- Clear, labeled axes and legends
Statistical Analysis (10%)
- Appropriate test selection
- Correct implementation
- Effect size reporting
- Proper interpretation
- Connection to hypotheses
Discussion (5%)
- Accurate interpretation of findings
- Integration with literature
- Thoughtful limitations
- Reasonable future directions
- Broader cognitive implications
3. Oral Presentation (15% - Finals Week)
Purpose: Authentic assessment verifying genuine understanding, especially important given potential AI assistance with coding
Evaluation Criteria:
Content Understanding (5%)
- Can explain concepts without reading from slides
- Demonstrates deep comprehension of analytical decisions
- Handles questions confidently and accurately
- Makes meaningful theoretical connections
Methodological Justification (5%)
- Articulates clear rationale for preprocessing choices
- Connects decisions to component properties
- Defends choices when challenged
- Shows understanding of trade-offs
Communication Quality (5%)
- Clear, engaging delivery
- Appropriate pacing and time management
- Effective visual aids
- Professional presentation skills
- Responds to questions thoughtfully
Critical Feature: Questions are designed to probe whether students truly understand their code and decisions. Example questions:
- "Why did you choose a 0.1 Hz high-pass filter instead of 0.5 Hz?"
- "What would happen to your component if you used a different baseline period?"
- "Can you explain what this line of code is doing and why it's necessary?"
- "How did you decide whether to use peak or mean amplitude?"
Students who cannot answer these questions, suggesting they don't understand their own work, receive significantly reduced grades regardless of Live Script quality.
4. Self-Reflection Essay (5% - Finals Week)
Purpose: Encourage metacognition and honest engagement with learning process
Evaluation:
- Graded generously on thoughtfulness and honesty rather than "correct" answers
- Students who genuinely engage with prompts receive full credit
- Looks for evidence of:
o Self-awareness about learning process
o Critical reflection on challenges
o Honest acknowledgment of assistance received
o Thoughtful consideration of future improvements
Not evaluated on: Whether they struggled or succeeded; whether they used AI or not; writing quality beyond basic clarity
5. Informal Assessment Throughout
Week 10 Proposal:
- Ungraded but provides early warning if students misunderstand assignment
- Instructor feedback guides appropriate scope and focus
Lab Session Participation:
- Monitored informally to identify students who need additional support
- Asking questions and helping peers is encouraged
Office Hours Engagement:
- Students who seek help proactively often perform better
- Provides insight into common struggles for teaching adjustments
Evidence of Goal Achievement
Students have met learning goals when they:
1. Technical Competency: Live Script contains complete, documented analysis pipeline that runs without errors and produces meaningful results
2. Theoretical Understanding: Written sections accurately explain ERP component properties, neural generators, and cognitive functions, integrating multiple research sources
3. Computational Thinking: Every preprocessing step includes sophisticated justification explaining why that step is necessary and why specific parameters were chosen
4. Critical Thinking: Discussion section thoughtfully addresses limitations, considers alternative interpretations, and suggests meaningful improvements
5. Communication Skills: Presentation clearly conveys complex methods and findings; student can answer probing questions demonstrating genuine understanding
6. Metacognition: Self-reflection shows awareness of learning process, honest engagement with challenges, and thoughtful consideration of growth
Red Flags for Insufficient Learning
Warning signs that student hasn't met goals:
- Cannot explain code during presentation
- Justifications are superficial ("I used this filter because the tutorial said to")
- Results are presented without critical interpretation
- Unable to answer basic questions about their component
- Self-reflection shows no engagement with learning process
- Code works perfectly but student can't explain why
These cases typically result in significantly reduced grades and follow-up conversations about academic integrity.
Grade Distribution (Typical)
In well-scaffolded implementation:
- Most students (60-70%) earn B+ to A- grades - demonstrating solid understanding with good execution
- Top students (15-20%) earn A/A+ - sophisticated justifications, exceptional communication, deep theoretical integration
- Struggling students (10-15%) earn B- to C+ - complete the work but with superficial understanding or significant gaps
- Very few students (<5%) fail - usually due to incomplete submission or inability to demonstrate understanding in presentation
How Assessment Design Addresses AI Use
The multi-modal assessment ensures students can't simply generate a Live Script with AI and submit it:
1. Presentation reveals understanding: Students must explain their work orally and answer spontaneous questions
2. Justifications require thinking: AI can generate code, but thoughtful, context-specific justifications require understanding
3. Self-reflection encourages honesty: Explicitly asking about AI use normalizes it while requiring verification
4. Peer review exposes gaps: Discussing work with classmates reveals whether students understand their own code
5. Draft feedback identifies issues early: Instructor can probe understanding before final grading
This design philosophy: AI assistance is acceptable, but understanding is required.

For Students
EEGLAB Documentation and Tutorials
- EEGLAB Tutorial: https://eeglab.org/tutorials/ Comprehensive official tutorials covering all major functions used in the project. Students should work through "Getting Started" and "Processing EEG Data" sections before beginning the project.
- EEGLAB Wiki: https://eeglab.org/ Reference documentation for specific functions. Especially useful for troubleshooting and understanding function parameters.

ERP Analysis Methods
- Luck, S. J. (2014). An Introduction to the Event-Related Potential Technique (2nd ed.). MIT Press. The definitive textbook on ERP methodology. Chapters 5-7 cover preprocessing in detail. Essential reading for understanding analytical decisions. Available through most university libraries.
- Kappenman, E. S., & Luck, S. J. (Eds.). (2012). The Oxford Handbook of Event-Related Potential Components. Oxford University Press. Comprehensive reference for specific ERP components. Each chapter covers one component's properties, methods, and cognitive significance. Students should read the chapter relevant to their chosen component.
Component-Specific Literature Provided to students based on their component choice. Example papers for common components:
P300:
- Polich, J. (2007). Updating P300: An integrative theory of P3a and P3b. Clinical Neurophysiology, 118(10), 2128-2148. Comprehensive review of P300 theory and methodology.
N400:
- Kutas, M., & Federmeier, K. D. (2011). Thirty years and counting: Finding meaning in the N400 component of the event-related brain potential (ERP). Annual Review of Psychology, 62, 621-647. Seminal review of N400 research and interpretation.
N170:
- Rossion, B., & Jacques, C. (2012). The N170: Understanding the time course of face perception in the human brain. In The Oxford Handbook of ERP Components (pp. 115-142). Definitive guide to N170 component properties and measurement.
Statistical Analysis
- Statistical Methods in ERP Research: Luck, S. J., & Gaspelin, N. (2017). How to get statistically significant effects in any ERP experiment (and why you shouldn't). Psychophysiology, 54(1), 146-157. Critical discussion of appropriate statistical practices in ERP research, addressing common pitfalls.

For Instructors
Course Development Resources
- Delorme, A., & Makeig, S. (2004). EEGLAB: An open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. Journal of Neuroscience Methods, 134(1), 9-21. Original EEGLAB paper describing toolbox architecture and functionality. Useful for understanding design philosophy.

Software Requirements
Essential:
- MATLAB (R2020b or later recommended) with academic license
- EEGLAB toolbox (version 2021.0 or later): Free download from https://eeglab.org/
- Signal Processing Toolbox (typically included in academic MATLAB licenses)
- Statistics and Machine Learning Toolbox (for statistical analyses)
- ERPLAB toolbox: https://erpinfo.org/erplab - Complementary to EEGLAB with additional ERP-specific functions