Innovative Research in Neural Processing

Pioneering hybrid approaches to BCI systems for enhanced accuracy and task performance through advanced signal processing and language model integration.

A device with a black interface, displaying green digital numbers and symbols, rests on a folded grey fabric surface. To its left, a coil of red and black cables is partially visible, and the background features green grass, indicating an outdoor setting.
A device with a black interface, displaying green digital numbers and symbols, rests on a folded grey fabric surface. To its left, a coil of red and black cables is partially visible, and the background features green grass, indicating an outdoor setting.

Neural Processing

Combining neural signals with language models for enhanced BCI performance.

A white Flick HAT device is placed on a wooden table. To the left is a hand, slightly out of focus, reaching towards the device. On the right, there are orange boxes with 'Flick' branding and text indicating 3D tracking and gestures for Raspberry Pi.
A white Flick HAT device is placed on a wooden table. To the left is a hand, slightly out of focus, reaching towards the device. On the right, there are orange boxes with 'Flick' branding and text indicating 3D tracking and gestures for Raspberry Pi.
Hybrid Approach

Our hybrid approach integrates neural signal processing and language models to improve BCI systems, focusing on error prediction and user satisfaction through controlled experimental evaluations.

A person's hand resting on a chair arm while wearing a blue clip device connected to a cable. Medical monitors with blurred displays are in the background, indicating a hospital or clinical setting. A red strap is visible on the chair arm, suggesting stability or restraint for a procedure.
A person's hand resting on a chair arm while wearing a blue clip device connected to a cable. Medical monitors with blurred displays are in the background, indicating a hospital or clinical setting. A red strap is visible on the chair arm, suggesting stability or restraint for a procedure.
Error Compensation

We develop a fine-tuned GPT-4 system to predict intended actions, reducing errors in BCI applications and enhancing overall user experience through rigorous testing and comparative analysis.

Innovative Research

Transforming neural signals into actionable insights through advanced technology.

This system significantly improved my task completion rates and reduced errors.

The hybrid approach effectively predicted my intended actions, enhancing user satisfaction and performance in various tasks beyond standard BCI systems.

Innovation

Combining neural signals with language model insights.

© 2025. All rights reserved.