Kandu Takes Brain-Computer Interface Stroke Rehab Into Patients' Homes

Kandu has built an at-home brain-computer interface for stroke rehabilitation, moving BCI-based therapy out of the clinic and into the living room — a shift that could redefine how hundreds of thousands of chronic stroke survivors access neuroplasticity-driven treatment.

The device, profiled by Citeline, takes what the company calls a "can-do" approach to post-stroke recovery. Rather than framing rehabilitation around deficits, the system is designed to reward what the patient's brain can still produce, using recorded neural signals to drive functional training. That design philosophy aligns with a growing body of evidence that brain-computer interfaces can harness intact cortical intention — motor imagery and residual motor activation — to retrain damaged motor networks.

The commercial logic is straightforward. Conventional BCI stroke rehabilitation has lived almost entirely in research laboratories and inpatient settings, where high-density scalp EEG, technical staff, and costly hardware limit session frequency. Home-based systems promise the opposite: many short sessions, sustained over weeks or months, which is precisely the dosing regimen that motor relearning studies suggest matters most.

For clinical neurophysiologists, the salient question is what the recording modality is and how signal quality holds up outside a controlled environment. Kandu's platform is a non-invasive, at-home system — meaning scalp-level recordings rather than intracranial electrodes or electrocorticography. At-home use carries familiar engineering constraints for any EEG reader: electrode impedance drift, motion artifact, ambient electrical noise, and the absence of a technologist to troubleshoot montage problems in real time. The company's ability to manage those variables in untrained hands will determine whether the device delivers interpretable neural signals or noise.

The regulatory status of the device is a key consideration for reading-room and monitoring-unit audiences weighing how to discuss it with patients. Companies in this space occupy a spectrum from FDA-cleared rehabilitation tools to investigational systems used only under study protocols, and the distinction matters for what clinicians can ethically recommend today. The Citeline profile positions Kandu as a commercial venture aimed at the stroke rehabilitation market, though clinicians should verify the device's specific clearance pathway and any published efficacy data before counseling patients.

The broader context lends the project weight. Stroke remains a leading cause of long-term disability, and a large share of survivors live with chronic upper-limb impairment that conventional therapy only partially addresses. BCI-based approaches — pairing motor-intention decoding with real-time feedback, often through exoskeletons, functional electrical stimulation, or on-screen avatars — have accumulated supportive evidence in small and mid-sized trials over the past decade. What those trials have generally lacked is scalability.

At-home deployment is the scalability argument in its purest form. If a patient can don a headset, run a session without supervision, and still generate training-relevant neural features, the effective "dose" of neurofeedback-mediated motor training could multiply several-fold compared with clinic visits two or three times per week.

The claims, for now, should stay tethered to what Kandu has actually demonstrated: a home-use BCI platform built around a capability-focused rehabilitation model. Peer-reviewed outcome data — ideally randomized, with standardized motor assessments such as the Fugl-Meyer scale and clearly reported EEG acquisition parameters — will determine whether the "can-do" framing translates into measurable motor recovery.

Kandu's bet is that the next phase of BCI stroke rehabilitation will be won not in the laboratory but in the patient's home, where frequency of use, not sophistication of hardware, drives cortical change.