Advancing computational neuroscience for content
Our research bridges the gap between neuroscience and creative technology.
Papers
ORCLE-nano: A small scale, 6-Stream Brain Encoder for Neural Prediction
kwale.ai Research
Architecture and evaluation of ORCLE-nano, an Omnimodal Response And Cortical Latent Encoder foundation model that predicts cortical activation from video content.
Neural Metrics for Content Evaluation: From ROI Aggregation to Marketing KPIs
kwale.ai Research
Mapping atlas-based ROI aggregations to 20 interpretable neural metrics for content scoring.
Closed-Loop Content Optimization via Neural Feedback
kwale.ai Research
Using neural scoring as a reward signal for iterative content generation and editing.
Key references
The studyforrest dataset
fMRI data from participants watching Forrest Gump -- foundational training data for brain encoding models.
Algonauts Project
Challenge for predicting brain responses to visual stimuli. Key benchmark for neural encoding models.
fsaverage5 Surface Template
FreeSurfer surface template with 10,242 vertices per hemisphere (20,484 total) used for cortical mapping.
Interested in collaborating? Get in touch.