Famous physicist Stephen Hawking, sadly departed in 2018, was also affected by ALS. Picture taken from http://www.nbcnews.com/id/45886738/ns/health-aging/t/stephen-hawking-how-has-he-survived-so-long/#.XZtMCSgzaUk
Brain-Computer Interfaces (BCIs) are strategies used to establish communication between the brain and hardware without the need for any effectors of the body (e.g.: muscles used for movement). Such interfaces can allow communication with and restoration of motor function of, people with disorders such as Amyotrophic Lateral Sclerosis (ALS), spinal cord injury or Locked-in Syndrome (LIS). As well as to allow better control of prosthetic limbs in patients with such prosthetics. While the prevalence of ALS may not be particularly high (there is a homogeneous rate across Europe of about 2,2 cases per 100.000 people [1]), the socioeconomic costs associated with it are significant because of its devastatingly debilitating symptoms. With Europe’s aging population the incidents of ALS have been predicted to increase by approximately 69% by 2040 [2], spelling an unprecedented demand for the development of these technologies.
Of the existing non-invasive modalities for imaging the human brain, EEG, MEG, fMRI and fNIRS have all been used in BCI research. Due to its relatively low cost, portability, safety and resilience to electrical noise, the latter has been gaining more and more users ever since its suitability for the purpose was proven [3]. To acquire a readout from the subject’s brain there have been various techniques recruited. Some of the most popular are: (a) Performing mental arithmetic while monitoring the PFC, and, (b) Using motor imagery while monitoring the activity of the motor cortex. These actions can be associated with one answer in a binary paradigm (e.g.: ‘yes’) while relaxation can signify the other (e.g.: ‘no’). A still from a participant undergoing one such experiment is shown below.
An obvious limitation of fNIRS as a tool for BCIs is its poor temporal resolution. Questions which are amenable to a binary answer are not usually very illuminating regarding one’s internal state, meaning that the paradigm needs to move fast if it is to facilitate meaningful communication. Similarly, if a speller is to be used to allow patients to construct sentences, alternative choices will be numerous and long waiting times needed for the haemodynamic response to develop may end up making the process daunting.
Figure was taken from the publication Borgheai et al., 2019. It shows the experimental design used for the selection of one letter in the speller created by the group. By using the matrix design the group severely cut down on the required time to choose one character.
In recent attempts to cut down on waiting times, Borgheai et al., 2019 [4], came up with an innovative design which makes use of a visual matrix of letters (see Figure below). Participants were asked to do some mental arithmetic task for 6 seconds when the line or column containing their letter of choice was illuminated. Even though the interval between letter choices was set to six seconds in order to increase the accuracy of the classification, authors reported a good accuracy (approximately 75%) using data from only the first two seconds of stimulation.
This study used a BCI system based solely on fNIRS. However, in many recent publications of the field where a hybrid BCI system (combined EEG and fNIRS) was used, there has been an enhanced classification accuracy reported [5]. This is perhaps not so surprising since EEG is a fundamentally different modality with strengths which complement those of fNIRS. In the future, combining these innovations may well provide a cutting edge for communication and rehabilitation strategies aimed at patients who suffer from conditions which impair the motor system.
References
1. Logroscino, G. et al. Incidence of amyotrophic lateral sclerosis in Europe. Journal of Neurology, Neurosurgery and Psychiatry 81, 385–390 (2010).
2. Arthur, K. C. et al. Projected increase in amyotrophic lateral sclerosis from 2015 to 2040. Nature Communications 7, (2016).
3. Coyle, S., Ward, T., Markham, C. & McDarby, G. On the suitability of near-infrared (NIR) systems for next-generation brain-computer interfaces. in Physiological Measurement 25, 815–822 (2004).
4. Borgheai, S. B., Abtahi, M., Mankodiya, K., McLinden, J. & Shahriari, Y. Towards a Single Trial fNIRS-based Brain-Computer Interface for Communication. in International IEEE/EMBS Conference on Neural Engineering, NER 2019-March, 1030–1033 (IEEE Computer Society, 2019).
5. Fazli, S. et al. Enhanced performance by a hybrid NIRS-EEG brain computer interface. NeuroImage 59, 519–529 (2012).
*Thumbnail picture was taken from https://aboveintelligent.com/brain-computer-interfaces-why-why-now-d02445090509