IEEE Systems, Man and Cybernetics Magazine - July 2020 - 17

focused on the machine-learning challenges of this new
framework. Efforts have mainly concentrated on the algorithmic modifications needed to solve the mathematical
optimization problem associated with the parameter estimation of each decoder type during real-time BMI operation (instead of the traditional batch and offline approach).
But, can subject learning happen under the dynamics generated by an adaptive BMI decoder? Because adaptive
BMIs inevitably result in the situation where a given neural
activity pattern can lead to different BMI outputs within
short amounts of time and thus to an unstable and confusing feedback, the concern has been sensibly raised [12],
[66] as to whether it is reasonable to expect subjects to
learn such an "ever-changing" task.
Given the absence of longitudinal studies of truly BMI
coadaptation in humans, the answer to this question
remains largely speculative. Nevertheless, gleaning evidence
from the collection of relevant literature converges toward
the following conclusion: BMI decoder adaptation during
closed-loop control should only be enabled in the beginning
of new BMI (training or operation) sessions until nonstationarity effects are alleviated and BMI performance restored.
Subsequently, stable, or at least only smoothly adapting,
decoders should be preferred so as to foster and exploit subject-learning capacities. Partial support for this view comes
from the few longitudinal studies with humans that have
showcased subject-learning effects with BMIs involving no
or only parsimonious decoder adaptation [18], [20], [57].
Collinger et al. [57] adopted the methodology of previous works on primates with daily, session-wise rerecalibration; however, parameter re-estimation happened only
in the beginning of sessions and decoders were kept fixed
for the remainder of each session, in full accordance with
our proposition. Perdikis et al. [20] modified the decoders
only if the new one outperformed the current decoder,
something that happened sporadically. McFarland et al.
[55] is the only case in which subject-learning effects seem
to accompany continuous decoder adaptation. Nevertheless, it is not clear whether the learning rate hyperparameter led to intensive or mild decoder modifications. In the
latter case, it can be assumed that an adaptive, but still
"approximately stable" decoder may not have disturbed
the subjects' learning efforts.
Several BMI studies on animals are particularly supportive of stable decoders' ability to foster subject learning
[3], [6]. In particular, Orsborn et al. [13] presented the only
study explicitly designed to answer the question of whether
consolidated subject learning is possible during BMI adaptation. The authors highlight the risk of the "moving target"
problem, commenting that the performance variability in
previous studies employing online or recurrent BMI adaptation may be due to the unsuitability of these approaches
for inducing permanent neuroplasticity and consolidated
skill formation, in spite of the average performance
improvement. Furthermore, they identify the study of the
interactions between subject and machine as key to
	

resolving these issues. They also suggest that decoder
adaptation is certainly beneficial only in terms of coping
with the nonstationarity of neural signals and the need to
track changes in neural ensembles (e.g., dying cells; the
equivalent in noninvasive BMI would be changes in the
optimal feature subset). Ultimately, they show the possibility of BMI skill acquisition with simultaneous decoder adaptation; however, the latter was "infrequent, minimal and
interspersed with long periods of fixed decoders." Hence,
continuous adaptation may not prevent subject learning as
long as the parameter update is mild, as shown in the earlier work of this and other groups [59], [81]. Lastly, the only
recent work that has attempted a generic mathematical
model of coadaptation [16] has also found in simulations
that, mild learning rates yield stable, converging systems
and should promote subject-learning effects.
In conclusion, we suggest that machine learning should
identify the optimal brain feature space, decode the brain
patterns therein, and track their shifts once nonstationarity effects occur; however, subject learning should be
responsible for increasing separability within this brain
manifold, a process that strongly relies on implicit/procedural mechanisms and requires substantial practice. We
consequently advocate to put more emphasis on exploring
novel paradigms that promote implicit subject learning of
BMI skills. This view on how to foster and unfold BMI
coadaptation, or any alternative one, can only be probed
through longitudinal and comprehensive experimental
assessments involving end users.
About the Authors
Serafeim Perdikis (serafeim.perdikis@essex.ac.uk) is
with the Brain-Computer Interfaces and Neural Engineering
Laboratory, School of Computer Science and Electronic Engineering, the University of Essex, Colchester, United Kingdom.
José del R. Millán (jose.millan@austin.utexas.edu) is
with the Department of Electrical and Computer Engineering and the Department of Neurology, the University of
Texas at Austin. He is a Fellow of the IEEE.
References
[1] U. Chaudhary, N. Birbaumer, and A. Ramos-Murguialday, "Brain-computer interfaces for communication and rehabilitation," Nat. Rev. Neurol., vol. 12, no. 9, pp.
513-525, 2016. doi: 10.1038/nrneurol.2016.113.
[2] N. Birbaumer et al., "A spelling device for the paralysed," Nature, vol. 398, no.
6725, pp. 297-298, 1999. doi: 10.1038/18581.
[3] J. M. Carmena et al., "Learning to control a brain-machine interface for reaching
and grasping by primates," PLoS Biol., vol. 1, no. 2, pp. 193-208, 2003. doi: 10.1371/
journal.pbio.0000042.
[4] C. Neuper and G. Pfurtscheller, "Neurofeedback training for BCI control," in
Brain-Computer Interfaces, B. Graimann, G. Pfurtscheller, and B. Allison, Eds. Berlin:
Springer-Verlag, 2009. pp. 65-78.
[5] B. Jarosiewicz, S. M. Chase, G. W. Fraser, M. Velliste, R. E. Kass, and A. B. Schwartz,
"Functional network reorganization during learning in a brain-computer interface
paradigm," Proc. Nat. Acad. Sci. U. S. A., vol. 105, no. 49, pp. 19,486-19,491, 2008. doi:
10.1073/pnas.0808113105

Ju ly 2020

IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE	

17



IEEE Systems, Man and Cybernetics Magazine - July 2020

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - July 2020

Contents
IEEE Systems, Man and Cybernetics Magazine - July 2020 - Cover1
IEEE Systems, Man and Cybernetics Magazine - July 2020 - Cover2
IEEE Systems, Man and Cybernetics Magazine - July 2020 - Contents
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 2
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 3
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 4
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 5
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 6
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 7
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 8
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 9
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 10
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 11
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 12
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 13
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 14
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 15
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 16
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 17
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 18
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 19
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 20
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 21
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 22
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 23
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 24
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 25
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 26
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 27
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 28
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 29
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 30
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 31
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 32
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 33
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 34
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 35
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 36
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 37
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 38
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 39
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 40
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 41
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 42
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 43
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 44
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 45
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 46
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 47
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 48
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 49
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 50
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 51
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 52
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 53
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 54
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 55
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 56
IEEE Systems, Man and Cybernetics Magazine - July 2020 - Cover3
IEEE Systems, Man and Cybernetics Magazine - July 2020 - Cover4
https://www.nxtbook.com/nxtbooks/ieee/smc_202110
https://www.nxtbook.com/nxtbooks/ieee/smc_202107
https://www.nxtbook.com/nxtbooks/ieee/smc_202104
https://www.nxtbook.com/nxtbooks/ieee/smc_202101
https://www.nxtbook.com/nxtbooks/ieee/smc_202010
https://www.nxtbook.com/nxtbooks/ieee/smc_202007
https://www.nxtbook.com/nxtbooks/ieee/smc_202004
https://www.nxtbook.com/nxtbooks/ieee/smc_202001
https://www.nxtbook.com/nxtbooks/ieee/smc_201910
https://www.nxtbook.com/nxtbooks/ieee/smc_201907
https://www.nxtbook.com/nxtbooks/ieee/smc_201904
https://www.nxtbook.com/nxtbooks/ieee/smc_201901
https://www.nxtbook.com/nxtbooks/ieee/smc_201810
https://www.nxtbook.com/nxtbooks/ieee/smc_201807
https://www.nxtbook.com/nxtbooks/ieee/smc_201804
https://www.nxtbook.com/nxtbooks/ieee/smc_201801
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_1017
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0717
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0417
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0117
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_1016
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0716
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0416
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0116
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_1015
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0715
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0415
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0115
https://www.nxtbookmedia.com