CogLab

Inspiration

As someone who has been in cognitive science and psychology labs for 6 years in college, I experienced the pain point of creating and building behavioral experiments. The opportunity is huge — universities spend over $7B a year on IT, and the global UX research software market is growing fast, from $400M in 2023 to a projected $1.7B by 2032. Together, that’s a ~$9B market we can capture by making experimentation as simple as drag-and-drop.

Pain points experienced:

High learning curve for tools like PsychoPy and jsPsych.
Experiment creation is time-consuming and tedious.
Short master’s programs create time pressure, limiting research scalability/efficiency.
Experimenter bias when conducting experiments.
Replication crisis - studies are difficult to replicate reliably. (63% of psychology studies not able to be replicated and produce the significant findings of the original studies.)

What it does

AI-powered experiment builder: generates behavioral experiments without coding.

Key features:

Drag and drop experiment builder, complete with consent, debrief, demographic questions, and tasks.
Supports complex designs, stimuli, and response logging.
Optional data analysis: cleans, structures, and visualizes results.
Easily replicable across participants and labs.

How we built it

Front-end: React, TypeScript, JavaScript
Back-end / Scripting: Python, Next.js, Vite
AI & Voice: OpenAI (image generation), DALL·E (images)
Workflow: Full-stack web app hosted on AWS EC2

Other tools: APIs integrated for stimuli generation and experiment management

Challenges we ran into

Handling timing and response recording for memory/recall tasks.
Ensuring unbiased instructions with voice AI. (future feature)
Designing a flexible system that supports multiple experiment types.
Balancing demo simplicity with showing the full experiment workflow.

Accomplishments that we're proud of

Built a fully functional AI experiment builder in a short hackathon timeframe.
Implemented consent, stimulus presentation, recall, and logging in a single workflow.
Added voice instructions to remove experimenter bias.
Thought through and designed the entire experiment process, from setup to data collection.
Demonstrated potential to automate data analysis and expand to UX research.

What we learned

Automating experiment creation reduces barriers for researchers without coding skills.
Voice AI can standardize instructions and improve reliability.
Even small design choices (timing, button hierarchy) impact usability and data quality.
Rapid prototyping highlights how AI can accelerate research pipelines.