The Time for Translation of Mobile Brain and Body Imaging to People With Stroke Is Now

While regaining walking and arm use are top priorities for individuals with stroke,1 there is a significant gap between interventions that show changes in clinical outcome measures and their translation to functional ability in the real-world. For example, clinician-reported assessments likely do not reflect the richness of naturalistic movement, and patient-reported outcome measures (PROMs) leave an incomplete picture of the level of function in stroke.2 Research that employs mobile brain/body imaging (MoBI)3 is uniquely positioned to fill this knowledge gap, as MoBI objectively measures the underlying neurophysiology linked to the behavior of functional, real-world movements that can supplement clinician-reported outcomes and PROMs. Thus, neurophysiological data that explain body movements derived from MoBI could identify which therapeutic interventions improve real-world activities. Here, we define MoBI as an experimental design where data from at least one neuroimaging technique and body behavioral measurement are used. For example, eye-tracking glasses could be paired with a mobile neuroimaging method such as electroencephalography (EEG) to understand the temporal coordination between the brain (patterns of cortical activity) and the body (eye movements) during walking (Figure). We argue that rehabilitation outcomes will be advanced by this technology in 2 ways: (1) MoBI can characterize the effectiveness of interventions in real-world environments, and (2) MoBI can guide evidence-based practice(s) in rehabilitation. One example of a mobile brain/body imaging experimental setup using electroencephalogram for neuroimaging, eye tracking for attention/navigation, heart rate sensor for stress monitoring, and inertial measurement units for movement tracking. One example of a mobile brain/body imaging experimental setup using electroencephalogram for neuroimaging, eye tracking for attention/navigation, heart rate sensor for stress monitoring, and inertial measurement units for movement tracking. Stroke is the leading cause of long-term disability in the world4 and leads to a low health-related quality of life.5 While rehabilitation methods exists, few are effective6 and the efficacy of others are still under debate.7 Therefore, there is a critical need for novel, effective interventions that improve motor function after stroke. It is crucial that clinical researchers investigate the efficacy of rehabilitation in an ecologically valid context, where gained knowledge is applicable to real-world situations and importantly, to individuals living in the community after a stroke. Despite innovative designs (eg, the use of foot pedals to understand gait,8 or a conveyor belt used to deliver tools to understand grasp9), magnetic resonance imaging (MRI) is one arm of research that cannot completely reflect certain real-world activities (eg, the use of a cane while walking). MRI studies require participants to remain motionless and rely on simulated or limited movements. Given the disconnect between functional MRI (fMRI) data and data collected in a real-world setting, it is possible that fMRI data do not provide accurate information regarding which rehabilitation strategies are effective.10 With technologies that can index brain activity in the real world, we believe it is time for researchers to incorporate MoBI into clinical studies. One promising MoBI technology is functional near-infrared spectroscopy (fNIRS), an imaging technique that uses light to measure changes in the oxygenation of hemoglobin, which is a proxy for energy expenditure in the brain. fNIRS can be used as a wearable device under ecologically valid conditions,11 has been validated as a reliable neuroimaging technique,12 and has high correspondence with fMRI findings.13,14 However, when brain activity was measured with fNIRS as healthy participants peeled an apple using a real knife, Okamoto et al observed prefrontal activity that was not evident when the same individuals pretended to peel an apple using a plastic knife during fMRI.15 While this study design has limitations (ie, increased prefrontal activation may be due to the increased attention a real knife requires to avoid cutting one’s self), this is one example of how brain activation during real-world tasks may differ from proxy tasks. It is possible that because many fMRI studies employ tasks and objects that are scanner safe, they have not captured brain activity that are crucial during real-world activities, a critique that also applies to most laboratory-based proxy tasks. As such, we have incomplete information as to how the brain supports function in real-world settings across all populations. If a MoBI study uncovers that a particular hand intervention in healthy adults engages the posterior parietal cortex, a region important for sensorimotor integration, in the future physical therapists could modify their intervention to personalize it for individuals with stroke in this brain region. For example, the physical therapist may want to include more somatosensory challenges like manipulating textured objects, as opposed to focusing on motor tasks alone. Thus, knowledge of brain activity could drive therapeutic choices. Further, how we conceptualize recovery after stroke, as well as measure the efficacy of interventions, may be incorrect as it applies to real-world settings. In addition to adopting ecologically valid study designs, it is essential that the field of rehabilitation pursue more objective dependent measures, such as MoBI, to determine the efficacy of interventions. Approximately 30% of individuals who have had a stroke experience cognitive impairment affecting domains such as attention, memory, and language (eg, difficulty with reading or writing).16 This challenges whether self-reported questionnaires, like PROMs, one approach clinical researchers use to assess the efficacy of rehabilitation, best reflect real-world activity. In addition, there is a limit of who can participate in questionnaire studies (eg, individuals who have Wernicke or Broca aphasia would not be able to participate); thus, questionnaire results may not be accurate when used with some individuals with stroke. MoBI is well positioned to provide objective evaluation of cognitive function without relying on questionnaires. Similarly, while motor assessments performed preintervention and postintervention provide researchers insight into the change of function over time, the movements performed in a laboratory often do not generalize to those performed in a complex environment (eg, toileting, reaching for a glass on the top shelf of a cabinet). One newer approach is the use of inertial measurement units to quantify movement time and kinematics including movement smoothness and hand trajectory. Inertial measurement units in conjunction with electromyography can distinguish an individual with a stroke from a healthy participant during arm movements,17 and a single unit placed in the pocket of a pair of pants can calculate the phases of gait.18 Another example of measuring real-world activity is with accelerometers. These devices can be placed around the wrist and ankle of individuals and continuously monitor activity for multiple days, extract features of movement, such as smoothness and strength,19 and produce data that are clinically meaningful.20 While we argue that it is imperative to use MoBI approaches in research in individuals with stroke, doing so requires addressing several challenges. One is participant safety. Executing movements like walking in the real world may result in injury. While we acknowledge that using MoBI in severely impaired patients or immediately following a stroke could be impractical, the crux of our argument is that an important first step is for the field to implement MoBI approaches into study designs. As such, researchers can use a MoBI design in individuals with stroke, which may offer an opportunity to study how brain and behavior interact in naturalistic environments. In addition, MoBI studies may be possible in more severely impaired individuals during cognitive or reaching tasks in naturalistic environments; this may accelerate research in individuals with severe motor or cognitive impairment, which is a need in the field.21 Understanding how the brain functions in naturalistic environments throughout recovery is the purpose of adopting MoBI approaches; however, an acceptable starting point is using MoBI in safe, controlled environments, such as rehabilitation research laboratories. Other challenges for implementing MoBI are device capabilities and size. While one of the benefits of MoBI devices are the relatively inexpensive and portable options, this also comes at the cost of reduced signal quality in comparison to medical-grade equipment used in laboratories. Subcortical structures cannot be reliably captured from MoBI devices such as electroencephalography or fNIRS. Moreover, the sensation of the equipment may draw attention to, or interfere with, a participant’s actions. This would defeat the purpose of MoBI. However, researchers and industry are working on smaller and less intrusive equipment, advanced signal processing, and machine-learning algorithms to overcome issues including signal noise, improper sensor placement, movement artifacts, and poor spatial resolution. At this time these are still obstacles every research study will have to address in their experimental designs. Owing to significant advancements in acute medical interventions, the number of individuals who survive a stroke are expected to double in the next 20 years.4 This demographic shift demands new, more effective rehabilitation approaches. However, our understanding of brain functions that support recovery and rehabilitation has been based on traditional neuroimaging methods that limit movement. MoBI affords us the opportunity to understand how a poststroke brain functions in a naturalistic environment. As we encourage researchers to adopt a MoBI approach in stroke populations, we propose the following guidelines. The engineering and physiological issues outlined in “Limitations of Mobile Brain/Body Imaging in Stroke Recovery Research” can be addressed in parallel to laboratory-based MoBI studies, with each technological advancement tested in the laboratory before being employed in real-world situations. Consequently, first we recommend basic studies that compare brain activity, eye movement, and muscle activity during complex movements to be executed in and outside the laboratory. While completing MoBI in a laboratory environment may make interpretation of results more straightforward, it likely produces results that are less reflective of how the brain and body behaves in the real world. Second, while we recommend participants execute movements that would reflect independent living with MoBI, we also recommend participants execute simple, restrictive movements that have been used during fMRI. This will allow researchers to compare previous results of brain activity and allow for the validation of data from fNIRS and electroencephalography. Third, we urge researchers to use MoBI to understand whether experimental interventions aimed at improving brain function (eg, noninvasive brain stimulation, exercise, blood flow restriction) in individuals with stroke are effective, how they compare to one another and to traditional therapies, and how they may change the brain and body over time. The penultimate goal of this work is to develop effective training interventions informed by mechanistic data that improve recovery and enhance health-related quality of life for individuals with stroke. Concept/idea/research design: B. Greeley, G. Hanada, L.A. Boyd, S. Peters Writing: B. Greeley, G. Hanada, L.A. Boyd, S. Peters Data collection: B. Greeley, G. Hanada, L.A. Boyd, S. Peters Data analysis: B. Greeley, G. Hanada, L.A. Boyd, S. Peters Project management: B. Greeley, G. Hanada, L.A. Boyd, S. Peters Fund procurement: B. Greeley, G. Hanada, L.A. Boyd, S. Peters Proving participants: B. Greeley, G. Hanada, L.A. Boyd, S. Peters Providing facilities/equipment: B. Greeley, G. Hanada, L.A. Boyd, S. Peters Providing institutional liaisons: B. Greeley, G. Hanada, L.A. Boyd, S. Peters Clerical/secretarial support: B. Greeley, G. Hanada, L.A. Boyd, S. Peters Consultation (including review of manuscript before submitting): B. Greeley, G. Hanada, L.A. Boyd, S. Peters All authors were responsible for the concept of the manuscript. B. Greeley and S. Peters were responsible for the initial draft of the manuscript, whereas all authors were involved in the subsequent edits. G. Hanada produced the figure. This work was supported by Canadian Institutes of Health Research Project Scheme 2016 (sponsor identifier: PJT-148535) and Fellowship (to S.P.) and the Michael Smith Foundation for Health Research Fellowship (to S.P.). The funders played no role in the writing of this point of view. The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.