Page 88 - GTM-3-1

Global Translational Medicine BCI-FES with static magnetic field in SCI

4.3. Interventions King et al. developed a NES-FES interface to walk over
39
4.3.1. Static magnetic field the ground for a duration of 30 sessions and evaluated
its performance in an individual with paraplegia due
In this study, there was no improvement in neuromuscular to SCI. This study investigated sessions in (i) screenings
condition (Section 4.1), possibly due to factors related to (BCI screening to determine whether one could control
the severity of the injury (injury time, extent, level, and the BCI in a virtual reality environment—four sessions
classification). The results observed in a human model total) and (ii) training (BCI training to learn to ambulate
(chronic condition) are different from those in animal within a virtual reality environment) using the attempt to
models in the acute phase. For instance, Bhattacharyya walk and relax as a control strategy, starting in the fifth
et al. exposed rats to a low-intensity magnetic field session and extending to the last. Their study concluded
34
(17.94 μT) for 2 h/day for 3 weeks after SCI. The results that the restoration of brain-controlled walking is feasible.
suggest that short-term exposure to a magnetic field on In the current study, the participants performed only three
locomotor behavior is beneficial because of the attenuation initial sessions to get acclimated, and despite having had a
of secondary damage and calcium ion-mediated relatively low performance in the training of the following
excitotoxicity. sessions, interventions with the sNES-sFES interface
Although the intervention in the present study was had already started. Thus, it is essential to invest in more
performed at the spinal cord level, studies using the SMF sessions to familiarize the participants before starting the
in the human cortex have shown that this intervention interventions, including avoiding anxiety due to their
modulates cortical excitability and improves the detection responsibility to master the interface.
35
of somatosensory stimuli when applied to the parietal 4.4. User learning curve
cortex. Another item that should be investigated in depth
36
is the ideal time for the application of the SMF. Despite 4.4.1. Classifier accuracy
the extended duration of SMF application in our study The AcCSP-LDA value in this study increased over the
(approximately 1 h), a study by Oliviero et al. showed that training sessions, reaching an average of 67%. Comparatively,
37
the application of transcranial SMF for up to 2 h is safe and King et al. obtained an average value of cross-correlation
39
does not appear to produce neural damage. between the instructor’s verbal commands and the response
38
Nakagawa and Nakazawa investigated the effects of of the sNES-sFES interface of 77%, reaching 93% between
trans-spinal SMF stimulation applied to the cervical spinal sessions 20 and 21. Therefore, this study demonstrated
cord on corticospinal excitability. They used a magnet that restoring brain-controlled walking in patients with
with a magnetic induction value of 0.45 T positioned paraplegia is feasible. Given this, although the average value
on the spinal cord level C8 of 24 healthy individuals for of the participants in the present study was 67%, in the last
15 min. The motor-evoked potentials of the first digital sessions, the values of cross-correlation were close to the
interosseous muscle were evaluated before, during, and average cited by King et al. Thus, it was found that the
39
after trans-spinal SMF stimulation. They showed that this learning effect interferes with the brain-machine domain,
intervention could reduce corticospinal tract excitability. and the user’s performance may have been influenced by
Therefore, researchers have suggested that trans-spinal the reduced number of sessions.
SMF stimulation can be a neuromodulatory tool, and
its suppressive effect can be in those with pathological 4.4.2. Hits and misses
hyperexcitability of the spinal neural network, similar The hits (intentional response) and errors (unintentional
to spasticity, a common condition in individuals with response) of AcCSP-LDA were tallied, and any
complete SCI. As in the present study, the participant unintentional response of the device controlled by the
presented a reduction in muscle tone, which can be participant was associated with the potential errors
explained both by the use of the NES-FES interface and found in the sNES. To avoid incorrect actions by BCI
40
also by the SMF. We cannot say whether this effect was and improve performance, Liu et al. proposed a new
41
potentiated by the combined use of these interventions, method based on an adaptive autoregressive model and a
and future investigations are necessary. typical spatial pattern for extracting error-related potential
features. They showed that the average precision and
4.3.2. Neuroelectrical signals and functional electrical false positive rate for detecting error-related potential
stimulation interface outperformed methods that use features extracted from a
Although the participant did not exhibit improvement single domain. This new method efficiently improves the
in his neuromuscular condition with the use of the NES- detection accuracy of error-related potentials and reduces
FES interface, it remains an assistive technology option. the number of false positives.

Volume 3 Issue 1 (2024) 8 https://doi.org/10.36922/gtm.2285