Creating an Engaging BCI/FES Therapy for Children With Perinatal Stroke Using Social Media

Creating an Engaging BCI/FES Therapy for Children With Perinatal Stroke Using Social Media

The goal of this trial is to evaluate the feasibility of a brain-computer interface controlled functional electrical stimulation (BCI-FES) social media integrated system for children with hemiparetic cerebral palsy . The main questions it aims to answer is:

(1) Is the social media BCI-FES system a feasible therapeutic intervention?

Perinatal strokes are focal neurovascular brain injuries that affect 5 million people worldwide and can lead to lifelong physical disabilities. It is the leading cause of hemiparetic cerebral palsy (HCP) which is characterized as motor dysfunction on one side of the body. Even modest improvements in a child's hand function can positively impact participation and affect quality of life, which is why there is a push for therapeutic interventions during earlier developmental windows.

Functional electrical stimulation (FES) stimulates muscle contraction during activity performance through low-intensity electrical currents. When applied to the affected limb, it can reduce spasticity-related symptoms and improve range of motion (ROM) and strength. When a patient's attempted movement is paired with FES, it triggers the cortical activation of sensorimotor areas, the degree of which has been associated with functional improvement. In stroke rehabilitation guidelines for adults, upper extremity FES receives the highest evidence-based recommendation. However, it has not been adequately studied in children, largely due to challenges with engagement, since high repetition counts are required to achieve meaningful functional gains. Brain Computer Interfaces (BCI) work by converting intentional brain activity into commands that can be used to control external devices. BCIs have been recently under investigation for their role in stroke rehabilitation.Specifically, BCI paired with FES has been explored as a novel way to drive task-related neuroplasticity to improve motor impairments over time in adults with stroke. Evidence suggests BCI-FES may be effective in adult patients with stroke-induced hemiparesis. Despite the evidence of enhanced neuroplasticity, the use of FES and BCI in children remains limited. One study suggested BCI-FES is feasible and well-tolerated in children with perinatal stroke, but scored poorly on enjoyability, suggesting that uptake and engagement in the therapy require improved system design. Which is why, the investigators created a novel and engaging BCI-FES therapy designed for youth with the aim of increasing long-term efficacy.

Methods:

A multi-disciplinary team of therapists, engineers, researchers, clinicians, and patient partners helped design a BCI FES therapy that is integrated with social media. The resulting application was named FlickTok, a novel BCI/FES social media therapy system. FlickTok integrates three interacting components: 1) an EEG headset, 2) the FES system, and 3) an application to coordinate the two components with social media interactions.

Patient Partner Engagement The investigators engaged three youth patient partners to design and build a BCI-FES integrated with social media to ensure research focus was on patient identified priorities. Before development began, the investigatorsworked with the patient partners to ensure that the system aligned with target users They helped us pilot test the system to evaluate and improve technical performance and usability. Our patient partners were also actively involved in the study design, helping to develop and conduct the qualitative interviews and supported results interpretation.

EEG:

An EEG headset and sensors were used to read and record the user's brain activity. A gel-based g.tec97 headset used an EEG montage at FC5, FC1, FCz, FC2, FC6, C5 C3, C1, Cz, C2, C4, C6, CP5, CP1, CP2, CP6 with a reference electrode placed at FPZ and a ground electrode on the participant's earlobe.3 Data was recorded using the g.USBamp amplifier and EEG was sampled at 256 Hz.

FES:

The FES component was driven by a computer and stimulates the desired muscles for the execution of the target functional movement. Working with an occupational therapist (OT), the investigators chose three functional movements to target: wrist extension, thumb abduction, and finger extension. A set of disposable "2 × 2" carbon rubber electrodes were applied to the targeted motor nerves on the forearm individualized for each participant. Electrode placement differed between participants based on individual anatomy, function, and target movement chosen. If a target movement could not be generated with FES, the participant was excluded from the study.

Muscle stimulation was delivered using the Neurotrac Dual Channel Transcutaneous Electrical Nerve Stimulation (TENS) and Neuromuscular Electrical Stimulation (NMES) device with a remote switch. FES parameters differed between participants to maximize tolerability and was adapted if the participant became fatigued. To facilitate connecting the FES to the BCI output, the investigators built a custom designed micro controller-based interface that connected to a USB port, enabling direct communication from the computer to control the FES stimulator.

Application:

The FlickTok application was structured in two parts: a Python server that coordinated EEG signals built in-house using our open-source BCI software package, BCI-Essentials system. The custom BCI was used to classify the EEG signal as either 'Action' or 'Rest'. The second part consisted of an Electron-based user-facing application that integrated social media content and FES stimulation with the BCI. YouTube Shorts and Instagram Reels were used during both the training of the BCI and the session itself to increase participant engagement and motivation. To personalize the sessions, each participant chose a specific type of content they wanted to engage with.

Calibration of the BCI system consisted of twenty repetitions using a 2-second on, 2-second off protocol, adapted to identify periods of intentional "action" (on-condition) from "rest" (off- condition). This binary classification utilized a Riemannian Geometry based motor imagery classifier.100 After the training calibration, the application selected 2 second segments of the EEG data and triggered the FES if it was classified as attempted movement and did nothing if it was classified as rest (participants had the choice to trigger it if desired). There was a 5 second latency or 'rest' period after a swipe. The social media application was displayed on a monitor and coloured icons that represented the attempted movement were integrated to improve understanding of the system.

Inclusion Criteria:

-The investigators recruited participants 10-25 years, with perinatal stroke and disabling hemiparetic cerebral palsy.

Exclusion Criteria:

• Exclusion criteria include severe intellectual disability, unstable epilepsy, and orthopaedic surgery in the upper limb or botulinum toxin injection in the previous 6 months.