Title : Mapping the neuroimmune–decoder divide in chronic brain–computer interfaces: A bibliometric analysis of structural inference gaps in biomarker stability
Abstract:
Implantable BCIs and closed-loop neuroprosthetic systems have shown clinical effectiveness in treating neurological conditions such as paralysis, amyotrophic lateral sclerosis (ALS), movement disorders, and epilepsy. Their ongoing performance depends on the sustained stability of neural biomarkers used for adaptive decoding, including cortical, and single-unit signals. However, chronic implantation frequently results in signal drift and fluctuations of other variables that compromise therapeutic and communicative outcomes. These instabilities are closely linked to the foreign body response at the electrode interface, altering the local recording environment and degrading biomarker fidelity. This is well-documented in device and interface research, but is often not adequately addressed in decoder-focused studies, creating a structural inference gap that negatively impacts the long-term reliability of neuroprosthetics. To explore this issue, this review investigates the extent to which neuroimmune and neurovascular interface biology is incorporated into chronic brain-computer interface (BCI) decoder research, and examines the implications of its limited integration on interpreting biomarker instability.
We conducted a PRISMA-ScR-guided science-mapping analysis of peer-reviewed literature indexed in the Web of Science Core Collection and Scopus (2010-2026), particularly the top 100 most-cited chronic BCI papers searched using the keywords “brain-computer interface,” “neuroimmune interface,” and “biomarker stability”. Using validated Boolean searches, we screened records according to predefined eligibility criteria and curated them in Excel and BibTeX for analysis using RStudio’s bibliometrix and VOSviewer. Co-citation, keyword co-occurrence, thematic evolution, and bibliographic coupling methods were employed to assess the structure of the field, especially the extent of siloing and cross-domain integration, with an emphasis on key bridging papers.
The analysis contained 542 unique author keywords, with a dominant focus on device-level and electrophysiological constructs. Keywords such as “brain stimulation” (n=43), “deep brain” (n=40), “Parkinson's disease” (n=33), “closed loop” (n=31), and “local field potentials” (n=27) reveal the field's emphasis on closed-loop neuromodulation and subthalamic deep brain stimulation. In contrast, terms related to neuroimmune and neurovascular processes, such as “neuroinflammation,” “glial scarring,” “foreign body response,” and “blood-brain barrier,” were largely absent from the dominant keyword tier. The term “electrode impedance” (n=16) emerged as the closest biological proxy, suggesting limited attention to the perielectrode microenvironment.
The most-cited papers were those affiliated with the University of Pittsburgh, the Pennsylvania Commonwealth System, and the University of California System. High-impact journals such as Nature Biomedical Engineering dominated the citation landscape with a strong focus on neuro-engineering of biomaterials. Citation patterns peaked between 2012 and 2014 before declining, pointing to field maturation without a thematic shift toward interface biology. These findings suggest that while device performance and electrophysiological biomarkers themselves remain critical to chronic BCI research, the neuroimmune and neurovascular factors affecting biomarker stability have received comparatively less attention.
To conclude, the chronic BCI literature is primarily concerned with device performance and optimization, while the neuroimmune and neurovascular microenvironment, the biological basis for long-term signal stability, remains underexplored, which may lead to an overemphasis on decoder limitations. Integrating neuroimmune determinants into decoding models could enhance the long-term reliability and translational potential of implanted neuroprosthetic systems.
