Knowledge Vault 3/59 - G.TEC BCI & Neurotechnology Spring School 2024 - Day 5
Clinical and neuroscientific applications of an
intracortical brain-computer interface system
Dan Rubin, Harvard Medical School, Massachusetts General Hospital (USA)
<Resume Image >

Concept Graph & Resume using Claude 3 Opus | Chat GPT4 | Llama 3:

graph LR classDef braingate fill:#f9d4d4, font-weight:bold, font-size:14px; classDef clinical fill:#d4f9d4, font-weight:bold, font-size:14px; classDef applications fill:#d4d4f9, font-weight:bold, font-size:14px; classDef neuroscience fill:#f9f9d4, font-weight:bold, font-size:14px; classDef future fill:#f9d4f9, font-weight:bold, font-size:14px; A[Dan Rubin] --> B[BrainGate: intracortical brain-computer interface. 1] A --> C[Safety trial: 68 device-related events. 2] C --> D[Seizures post-implantation, protocol updated. 3] A --> E[BrainGate restores communication, mobility, independence. 4] E --> F[Typing at 40 characters/min with ALS. 5] E --> G[Integration with tablets, apps, environment. 6] E --> H[Handwriting decoded, 90 characters/min. 7] E --> I[Speech decoded, 60-80 words/min with ALS. 8] E --> J[Robotic arms, FES tested, some limitations. 9] E --> K[Soft robotic gloves for dexterous tasks. 10] E --> L[Wireless transmitters, BrainGate Home app. 11] A --> M[Middle frontal gyrus integrates auditory cues. 12] A --> N[Overnight task-related replay in motor cortex. 13] N --> O[Replay during slow-wave sleep, cortical ripples. 14] N --> P[Replay underlies consolidation, enhances performance. 15] A --> Q[BrainGate advances function and neuroscience. 16] A --> R[Multidisciplinary teams, dedicated participants. 17] A --> S[Pre-surgical imaging maps cortical areas. 18] A --> T[Sparse sleep activity challenges replay identification. 19] A --> U[Expanding speech decoder to other languages. 20] A --> V[Improving language model, home use, wireless. 21] A --> W[Integration with assistive tech, smart homes. 22] class A,B braingate; class C,D clinical; class E,F,G,H,I,J,K,L applications; class M,N,O,P,S,T neuroscience; class Q,R,U,V,W future;

Resume:

1.- BrainGate is an intracortical brain-computer interface system that has shown promising clinical and neuroscientific applications for people with paralysis.

2.- In a safety trial with 14 participants, there were 68 device-related adverse events, mostly minor, with no intracranial infections.

3.- Two participants had seizures post-implantation, likely due to prior brain injury and lack of anti-seizure prophylaxis, which was added to the protocol.

4.- BrainGate aims to restore communication, mobility, and functional independence for people with paralysis, going beyond current assistive technologies.

5.- Using 2D cursor control and virtual keyboards, BrainGate enabled typing at 40 characters per minute in a participant with ALS.

6.- BrainGate was integrated with commercial tablets and apps, allowing participants with ALS to write emails, search YouTube, and control their environment.

7.- Decoding attempted handwriting from motor cortex activity allowed a participant with spinal cord injury to write at 90 characters per minute.

8.- Decoding speech from motor cortex achieved communication rates of 60-80 words per minute in participants with anarthria due to ALS.

9.- Robotic arms and functional electrical stimulation were tested for restoring reach and grasp, though with some limitations in practicality and applicability.

10.- Soft robotic gloves controlled by BrainGate enabled a participant with spinal cord injury to perform dexterous hand movements and functional tasks.

11.- Wireless transmitters and a "BrainGate Home" app aimed to provide more independence in controlling personal devices and smart home environments.

12.- Decoding from middle frontal gyrus neurons showed their involvement in integrating auditory cues into movement planning, firing prior to motor execution.

13.- Overnight replay of task-related neural activity patterns was observed in human motor cortex during sleep, analogous to hippocampal replay in rodents.

14.- The replayed activity corresponded to a cursor moving through a recently practiced sequence, occurring preferentially during slow-wave sleep and cortical ripples.

15.- This neural replay likely underlies motor memory consolidation, as post-sleep task performance was enhanced. Further work explores ripples' role in inter-areal communication.

16.- BrainGate has the potential to restore meaningful function for people with communication and mobility impairments while advancing fundamental neuroscience.

17.- The work relies on multidisciplinary teams at several institutions and the dedication of research participants volunteering to advance the technology.

18.- Pre-surgical MRI, fMRI, and CT are used to map cortical structures and identify functionally relevant areas for implanting microelectrode arrays.

19.- Sparsity of neural activity during sleep poses challenges for identifying relevant replay events in large datasets; methods involve correlating with awake patterns.

20.- Expanding the BrainGate speech decoder to other languages would require adapting the phoneme set, language model, and participant population.

21.- Future priorities include improving the language model, automating and streamlining the system for independent home use, and reducing reliance on cables/pedestals.

22.- Integration with other assistive technologies and smart home devices via the BrainGate Home app is another key goal to enhance users' independence.

Knowledge Vault built byDavid Vivancos 2024