Neural NoiseBCI Weekly Brief (week of 2026-03-02)
Science Corp’s $230M raise ahead of an FDA decision on PRIMA dominates the industry signal this week. Research highlights include direct BCI decoding work from the Ramsey lab, a new MI-BCI decoding architecture, a rapid-fabrication subdural electrode array, and a Neuropixels real-time toolbox. Transcranial stimulation studies (tACS, TMS-EEG) round out the methods coverage.
STAT+: Science Corp. raises $230 million to bring retinal implant to Americans
STAT News
Published: 2026-03-05T19:01:41+00:00
Tags: science-corp, sensory-neuroprosthetics, PRIMA, FDA, funding, tier-1
Science Corp raises $230M while awaiting FDA decision on PRIMA wireless retinal implant. Major sensory-neuroprosthetics funding milestone FDA outcome could set precedent for visual BCI market entry. STAT is a top-tier health/biotech source.
- Science Corporation raised $230 million in a Series C funding round, bringing total funding to $490 million since its 2021 founding by Max Hodak.
- Investors include Lightspeed Venture Partners, Khosla Ventures, Y Combinator, Quiet Capital, and IQT (CIA’s investment arm).
- The company’s PRIMA wireless retinal implant, paired with near-infrared light glasses, restored central vision and improved visual acuity in late-stage macular degeneration patients after 12 months.
- Expected to launch in Europe later this year, pending FDA decision.
- Reported by O. Rose Broderick, who covers health policies and technologies for people with disabilities.
Science Corp. raises $230M as it races to bring its brain implant to market
TechCrunch - Biotech & Health
Published: 2026-03-05T14:00:00+00:00
Tags: science-corp, sensory-neuroprosthetics, funding, valuation, tier-1
Adds $1.25B post-money valuation detail to Science Corp raise. Confirms unicorn status for a sensory neuroprosthetics company pre-FDA decision, signaling strong investor confidence in the neural-implant sector.
- Raised $230M in Series C funding, bringing total funding to $490M and a post-money valuation of $1.5 billion.
- PRIMA: Brain-computer interface chip (smaller than a grain of rice) implanted in the eye, paired with camera-equipped glasses, to restore vision for advanced macular degeneration patients.
- Acquired PRIMA technology from Pixium Vision (2024) and refined it.
- Clinical trials (47 patients in Europe & U.S.): 80% showed meaningful improvement in visual acuity, enabling reading letters, numbers, and words.
- First documented restoration of fluent reading ability in blind patients.
- Regulatory Path: CE mark application submitted to the EU; approval expected mid-2026 (first market: Germany). Ongoing FDA discussions in the U.S.
- Expanding PRIMA trials to include Stargardt disease and retinitis pigmentosa.
- Developing a biohybrid neural interface with engineered neurons for brain surface integration.
- Launched Vessel, an organ preservation platform for miniaturized perfusion technology.
- Investors include Lightspeed Venture Partners, Khosla Ventures, Y Combinator, Quiet Capital, and IQT.
- Founded by Max Hodak (co-founder/former president of Neuralink); currently employs 150 people.
Dynamic graph based attention spectral network for motor imagery-brain computer interface
Frontiers in Human Neuroscience
Published: 2026-03-04T00:00:00+00:00
Tags: EEG-based-BCI, motor-imagery, neural-decoding, methods, tier-1
Proposes a graph-attention spectral network for MI-EEG decoding that models cross-frequency coupling and dynamic brain network organization. Directly targets BCI decoder performance—a core bottleneck for non-invasive MI-BCIs.
- Context: Motor imagery-based BCIs (MI-BCI) are increasingly used in neurorehabilitation, but current MI-EEG decoding algorithms often overlook the brain’s complex network organization and cross-frequency coupling (CFC) during motor imagery.
- Problem: Existing methods rarely account for how temporal dynamics across different MI stages affect decoding performance.
- Solution: The Dynamic Spectral-Spatial Interaction Convolution Neural Network (DSSICNN) is proposed. It is a parameter-efficient framework that jointly extracts temporal-spectral-spatial features.
- Innovations:
- Dual-branch architecture for Euclidean and non-Euclidean spatial representations.
- CFC-inspired attention module and attention mechanism for MI stage contributions.
- Results: DSSICNN outperforms state-of-the-art (SOTA) in both session-dependent and session-independent settings on public datasets.
- Impact: Offers design insights for GNN-based MI-EEG decoding and a network neuroscience perspective on MI mechanisms.
- Performance: DSSICNN achieves substantial improvements over SOTA, demonstrating its empirical advantages.
- Design Insights: Provides a framework for developing GNN-based MI-EEG decoding algorithms.
- Neurophysiological Understanding: Offers a network neuroscience-inspired perspective for interpreting the mechanisms underlying motor imagery.
A Novel Rapidly Manufacturable Flexible Subdural Electrode Array for Intraoperative Mapping of Cortical Activity
bioRxiv Neuroscience
Published: 2026-03-07T00:00:00+00:00
Tags: ECoG, neural-recording, electrode-array, biomedical-engineering, tier-1
Fast, low-cost PDMS/gold subdural electrode fabrication via laser cutter—bypasses cleanroom. Directly relevant to ECoG and intraoperative neural recording could accelerate custom array prototyping for BCI and surgical mapping labs.
- Objective: Develop a fast, low-cost method for fabricating flexible, biocompatible neural interfaces for intraoperative monitoring and chronic neural recordings.
- Materials: Polydimethylsiloxane (PDMS) substrate, gold conductive layer.
- Fabrication: Laser cutter used for mask generation and direct patterning of metal traces on PDMS, achieving up to 30 µm resolution.
- Interface: Detachable design for reliable connectivity during testing.
- Electrochemical: Ohmic behavior confirmed.
- Mechanical: Stable conductivity after 50 bending cycles, with less than 10% degradation.
- Electrochemical Impedance Spectroscopy (EIS): Validated electrode viability for physiological signal recording.
- Testing: Simultaneous recordings of local field potentials (LFPs) and electrocorticography (ECoG) in rat somatosensory cortex.
- Results: Flexible subdural array signals showed statistically significant (p < 0.001) median cross-correlation of 0.35 with LFPs recorded by industrial electrodes at 600–800 µm depth.
- Conclusion: Presents a robust, accessible approach for producing customizable, functional neural interfaces, suitable for rapid iteration in research and clinical settings.
OP-GLX: A MATLAB toolbox for online processing and plotting of Neuropixels data acquired with SpikeGLX
bioRxiv Neuroscience
Published: 2026-03-06T00:00:00+00:00
Tags: neural-data-analysis, Neuropixels, electrophysiology, tools, tier-1
Real-time processing toolbox for Neuropixels probes fills a gap in the neural data analysis pipeline. Useful for labs doing high-density electrophysiology who need online visualization alongside SpikeGLX acquisition.
- Purpose: Facilitates online processing and visualization of large-scale neural data, addressing a critical need in neuroscience and neural engineering.
- Context: Advances in recording technologies (e.g., Neuropixels probes) enable streaming from hundreds of electrodes, but real-time data handling remains challenging.
- Most existing software (e.g., SpikeGLX) focuses on acquisition stability, leaving processing and visualization for offline use.
- Compatibility: Operates alongside SpikeGLX to enhance data fetching, processing, and visualization.
- Processing Capabilities:
- Spike detection
- Time-binned firing rate computation
- Spike waveform plotting
- Principal component analysis (PCA)
- User Interface: Native GUI for intuitive, customizable interaction during experiments.
- Performance: Supports real-time operation, with no SpikeGLX stream buffer fetch errors in testing across various acquisition settings.
- Impact: Aims to improve researchers’ ability to visualize and interpret data during live neuroscience experiments.
Gamma tACS over the prefrontal and parietal cortices enhances episodic memory performance
Frontiers in Human Neuroscience
Published: 2026-03-02T00:00:00+00:00
Tags: tACS, transcranial-stimulation, neuromodulation, EEG, tier-1
Sham-controlled study shows gamma-tACS to PFC and PPC modulates episodic memory encoding and retrieval. Adds controlled evidence to the non-invasive neuromodulation literature relevant for closed-loop tACS and neurofeedback applications.
- Background: Episodic memory is crucial for daily functioning and is vulnerable to aging and neurological disorders; gamma-frequency tACS is a non-invasive method proposed to modulate memory-related brain activity.
- Objective: Assess whether gamma tACS (60 Hz) applied to the left prefrontal cortex (PFC) and posterior parietal cortex (PPC) during memory tasks improves episodic memory in healthy young adults.
- Methods:
- Participants: 51 right-handed, healthy adults (mean age: 20.9 years).
- Groups: Two-site (PFC–PPC), single-site (PFC), or sham stimulation (each n=17).
- Procedure: Verbal recognition task over three sessions (Days 1, 2, and 7); 60 Hz tACS (1.5 mA) during encoding and retrieval.
- Primary outcome: Discrimination index (d-prime) on Day 7.
- Results:
- Significant time and time-by-group effects for accuracy and d-prime.
- PFC–PPC group showed higher d-prime than sham on Days 2 and 7, with medium-to-large effect sizes.
- PFC group had numerically higher d-prime than sham on Day 7, but no robust statistical difference between active groups.
- Conclusions: Gamma tACS may enhance episodic memory discrimination at delayed time points; replication and neurophysiological studies are needed.
Distinct beta burst motifs exhibit opposing error relationships during motor adaptation
bioRxiv Neuroscience
Published: 2026-03-06T00:00:00+00:00
Tags: electrophysiology, MEG, motor-control, neural-signal-processing, tier-2
High-density MEG reveals distinct beta burst waveform motifs with separable computational roles during motor adaptation. Understanding burst heterogeneity matters for motor-BCI decoders that use beta features for movement prediction.
- Beta-band activity (13–30 Hz) is linked to human movement, but its functional role remains unclear.
- Beta activity consists of transient bursts with diverse waveforms, not just sustained oscillations.
- A study using high-density MEG recorded brain activity during a visuomotor rotation task under implicit and explicit learning conditions.
- Traditional beta metrics (power and burst rate) varied with context but did not explain trial-by-trial behavior.
- Analyzing burst waveforms revealed distinct burst types with unique dynamics and context-dependent changes.
- After movement, different burst subtypes showed opposing relationships with behavioral error: some decreased, others increased as error grew.
- Results indicate beta activity is made up of distinct events with specific computational roles, and waveform diversity is key to understanding its function in adaptive behavior.
Dynamic causal modelling for functional near-infrared spectroscopy using spatial priors derived from diffuse optical tomography
NeuroImage
Tags: fNIRS, methods, neuroimaging, tier-2
Advances fNIRS analysis by integrating diffuse optical tomography-derived spatial priors into dynamic causal modelling. Strengthens fNIRS as a non-invasive BCI-adjacent imaging modality with better source localization.
- Study Focus: Introduces dynamic causal modelling (DCM) for fNIRS using source-level brain regions estimated via diffuse optical tomography (DOT), moving beyond sensor-level analysis.
- Data: Uses experimental fNIRS data from 104 participants performing a Go/No-Go response inhibition task.
- Key Finding: DCM models with DOT-informed neuronal source locations show better model evidence than those using sensor-level locations.
- Neural Insight: Demonstrates suppressive (inhibitory) effects from the right inferior frontal gyrus (rIFG) to motor regions during response inhibition, though subcortical pathways are not directly modeled.
- Methodological Advantage: Combines DOT and SPM for depth-dependent source localization, improving causal connectivity inference from fNIRS data.
- Application: The portable nature of fNIRS allows this approach to be used in real-world, naturalistic settings for studying brain network modulation.
Reduced alpha-band phase coherence and cortical complexity in fibromyalgia: A tms-eeg exploratory study
Clinical Neurophysiology
Tags: TMS-EEG, clinical-neurophysiology, transcranial-stimulation, tier-2
Combines TMS-EEG to probe cortical excitability and complexity in fibromyalgia. Methodologically relevant for clinical neurophysiology and TMS-EEG pipeline development, though the clinical focus limits direct BCI application.
- Key Findings:
- Individuals with fibromyalgia (FM) exhibit reduced alpha-band phase synchronization (ITC) in parietal-occipital brain regions compared to healthy controls.
- FM patients show a diminished capacity to generate complex oscillatory brain activity (PCIst) after TMS stimulation.
- Reduced cortical connectivity and complexity in FM are linked to higher pain interference and less pain relief.
- Study Context:
- The study used TMS-EEG to measure cortical excitability and connectivity in FM patients (n=18) vs. controls (n=15).
- Abnormalities in brain connectivity were correlated with pain intensity, mood, and quality of life in FM.
- Implications: TMS-EEG may offer mechanistic insights and therapeutic targets for fibromyalgia.
Functional near-infrared spectroscopy identifies neural biomarkers of burnout in active-duty Police officers
Nature (Neuroscience subject)
Published: 2026-03-07T00:00:00+00:00
Tags: fNIRS, neuroimaging, biomarkers, tier-2
Uses fNIRS to identify prefrontal neural biomarkers of occupational burnout. Demonstrates fNIRS as a field-deployable neural sensing tool, though the application is clinical/behavioral rather than BCI.
- Burnout prevalence: Recognized by WHO, widespread across professions, and a significant global economic burden.
- Assessment challenge: Current methods rely on subjective scales, lacking objective reliability.
- Study objective: Develop an objective burnout assessment system using fNIRS and machine learning.
- Participants: 33 active police officers.
- Method: fNIRS measured prefrontal lobe hemoglobin changes during mental arithmetic and verbal fluency tasks; features extracted and labeled with burnout scale scores.
- Model: Support vector machine algorithm achieved 91.3% training accuracy and 90.0% test accuracy.
- Conclusion: fNIRS + machine learning shows promise for objective burnout assessment.
Focal Transcranial Magnetic Stimulation of the Rat Anterior Cingulate Cortex Inhibits Incubation of Opioid Craving after Voluntary Abstinence
bioRxiv Neuroscience
Published: 2026-03-06T00:00:00+00:00
Tags: neuromodulation, TMS, transcranial-stimulation, tier-2
Novel focal TMS system delivering theta burst stimulation with fMRI readout in rats. Advances mechanistic understanding of TMS engagement in the brain—relevant for translational neuromodulation, though the addiction focus limits direct BCI relevance.
- Challenge: Relapse is a major obstacle in opioid addiction treatment, with limited understanding of how neuromodulation therapies engage brain circuits.
- Model: Rats trained to self-administer oxycodone for 14 days, followed by 13 days of enforced abstinence using an electric barrier.
- Intervention: Daily high-density theta burst stimulation (hdTBS) or sham targeting the anterior cingulate cortex (ACC) for 7 days before testing.
- Assessment: Relapse-like behavior and ACC functional connectivity measured during early and late abstinence using resting-state fMRI.
- Sham Group: Showed increased oxycodone seeking over time (incubation of craving) and reduced ACC connectivity.
- hdTBS Group: Prevented craving incubation and restored ACC connectivity with the dorsal and ventral striatum.
- Mechanism: Effects of stimulation aligned with regions receiving dense ACC projections, suggesting projection architecture drives network modulation.
- Implication: ACC-centered circuits are viable targets for TMS-based interventions to reduce opioid relapse.