Clinical safety assessment of implanted brain-computer interfaces: current evidence and a pathway to standardization
Abstract
Implanted brain-computer interfaces are beginning to move beyond short laboratory sessions into long-term and home use. The evidence used to judge clinical safety remains fragmented. This review examines penetrating intracortical, cortical-surface (including subpial) and cerebral-venous systems, with stimulation considered as an additional exposure. Published studies report perioperative neurological and wound complications, vascular patency, stimulation-related effects, changes in devices and materials, and loss of functions on which participants may depend. Comparison across architectures remains difficult because cohorts are small and sometimes overlap, exposure denominators vary, long-term surveillance is incomplete, and retrieval data are limited. Chronic safety cannot be inferred from one test. Clinical review, problem-directed imaging, device interrogation and participant-level outcomes capture different parts of the safety record and need to be read together over time. We use this evidence to set out a lifecycle framework spanning the preimplant baseline, implantation, recovery, chronic use, new findings and end-of-use. It combines a small common dataset with architecture-specific modules and an event-triggered route for confirmation, attribution, escalation and outcome recording. Before this approach can inform formal guidance, it requires prospective multicentre testing, independent adjudication, input from participants and caregivers, registry follow-up and periodic revision. The near-term task is to make safety evidence comparable and clinically useful without treating any implanted BCI architecture as generally safe.
Related
- NeuroAtlas: Benchmarking Foundation Models for Clinical EEG and Brain-Computer Interfaces
- Neural signatures of glioma progression using implanted neural interfaces
- Human learning of noninvasive brain–computer interfaces via manifold geometry
- Real-time noise cancellation for EEG monitoring and brain-computer interfaces
- Large language models integrated into brain–computer interfaces for communication and control: a systematic review
BibTeX
@article{xing2026clinical,
title = {{Clinical safety assessment of implanted brain-computer interfaces: current evidence and a pathway to standardization}},
author = {Baozhou Xing and Li Zhang and Wen-Jun Tu},
journal = {Frontiers in Neuroscience},
year = {2026},
month = oct,
doi = {10.3389/fnins.2026.1994636},
url = {https://www.frontiersin.org/articles/10.3389/fnins.2026.1994636},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Clinical safety assessment of implanted brain-computer interfaces: current evidence and a pathway to standardization
AU - Baozhou Xing
AU - Li Zhang
AU - Wen-Jun Tu
PY - 2026
DA - 2026/10/28
JO - Frontiers in Neuroscience
DO - 10.3389/fnins.2026.1994636
UR - https://www.frontiersin.org/articles/10.3389/fnins.2026.1994636
AB - Implanted brain-computer interfaces are beginning to move beyond short laboratory sessions into long-term and home use. The evidence used to judge clinical safety remains fragmented. This review examines penetrating intracortical, cortical-surface (including subpial) and cerebral-venous systems, with stimulation considered as an additional exposure. Published studies report perioperative neurological and wound complications, vascular patency, stimulation-related effects, changes in devices and materials, and loss of functions on which participants may depend. Comparison across architectures remains difficult because cohorts are small and sometimes overlap, exposure denominators vary, long-term surveillance is incomplete, and retrieval data are limited. Chronic safety cannot be inferred from one test. Clinical review, problem-directed imaging, device interrogation and participant-level outcomes capture different parts of the safety record and need to be read together over time. We use this evidence to set out a lifecycle framework spanning the preimplant baseline, implantation, recovery, chronic use, new findings and end-of-use. It combines a small common dataset with architecture-specific modules and an event-triggered route for confirmation, attribution, escalation and outcome recording. Before this approach can inform formal guidance, it requires prospective multicentre testing, independent adjudication, input from participants and caregivers, registry follow-up and periodic revision. The near-term task is to make safety evidence comparable and clinically useful without treating any implanted BCI architecture as generally safe.
SN - 1662-453X
ER -