ORIGINAL RESEARCH published: 07 July 2020 doi: 10.3389/fnins.2020.00582 EEG Biomarkers Related With the Functional State of Stroke Patients Marc Sebastián-Romagosa 1,2* , Esther Udina 1 , Rupert Ortner 2 , Josep Dinarès-Ferran 2,3 , Woosang Cho 4 , Nensi Murovec 4 , Clara Matencio-Peralba 2 , Sebastian Sieghartsleitner 4 , Brendan Z. Allison 5 and Christoph Guger 2,4 1 Department of Physiology, Universitat Autònoma de Barcelona, Barcelona, Spain, 2 g.tec Medical Engineering Spain SL, Barcelona, Spain, 3 Data and Signal Processing Research Group, Department of Engineering, University of Vic - Central University of Catalonia, Vic, Spain, 4 g.tec Medical Engineering GmbH, Schiedlberg, Austria, 5 Department of Cognitive Science, University of California at San Diego, La Jolla, CA, United States Introduction: Recent studies explored promising new quantitative methods to analyze electroencephalography (EEG) signals. This paper analyzes the correlation of two EEG parameters, Brain Symmetry Index (BSI) and Laterality Coefficient (LC), with established functional scales for the stroke assessment. Methods: Thirty-two healthy subjects and thirty-six stroke patients with upper extremity hemiparesis were recruited for this study. The stroke patients where subdivided in three groups according to the stroke location: Cortical, Subcortical, and Cortical + Subcortical. The participants performed assessment visits to record the Edited by: EEG in the resting state and perform functional tests using rehabilitation scales. Then, Ioan Opris, stroke patients performed 25 sessions using a motor-imagery based Brain Computer University of Miami, United States Interface system (BCI). BSI was calculated with the EEG data in resting state and LC Reviewed by: was calculated with the Event-Related Synchronization maps. Riccardo Iandolo, Italian Institute of Technology, Italy Results: The results of this study demonstrated significant differences in the BSI Solaiman Shokur, École Polytechnique Fédérale between the healthy group and Subcortical group (P = 0.001), and also between the de Lausanne, Switzerland healthy and Cortical+Subcortical group (P = 0.019). No significant differences were *Correspondence: found between the healthy group and the Cortical group (P = 0.505). Furthermore, Marc Sebastián-Romagosa sebastian@gtec.at; the BSI analysis in the healthy group based on gender showed statistical differences marc.sebastian@e-campus.uab.cat (P = 0.027). In the stroke group, the correlation between the BSI and the functional state of the upper extremity assessed by Fugl-Meyer Assessment (FMA) was also significant, Specialty section: This article was submitted to ρ = −0.430 and P = 0.046. The correlation between the BSI and the FMA-Lower Neural Technology, extremity was not significant (ρ = −0.063, P = 0.852). Similarly, the LC calculated in the a section of the journal alpha band has significative correlation with FMA of upper extremity (ρ = −0.623 and Frontiers in Neuroscience P < 0.001) and FMA of lower extremity (ρ = −0.509 and P = 0.026). Other important Received: 30 January 2020 Accepted: 12 May 2020 significant correlations between LC and functional scales were observed. In addition, Published: 07 July 2020 the patients showed an improvement in the FMA-upper extremity after the BCI therapy Citation: (1FMA = 1 median [IQR: 0–8], P = 0.002). Sebastián-Romagosa M, Udina E, Ortner R, Dinarès-Ferran J, Cho W, Conclusion: The quantitative EEG tools used here may help support our understanding Murovec N, Matencio-Peralba C, of stroke and how the brain changes during rehabilitation therapy. These tools can help Sieghartsleitner S, Allison BZ and Guger C (2020) EEG Biomarkers identify changes in EEG biomarkers and parameters during therapy that might lead to Related With the Functional State improved therapy methods and functional prognoses. of Stroke Patients. Front. Neurosci. 14:582. Keywords: brain-computer interface, motor imagery, EEG, rehabilitation, Brain Symmetry Index, laterality doi: 10.3389/fnins.2020.00582 coefficient Frontiers in Neuroscience | www.frontiersin.org 1 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation INTRODUCTION During the MI tasks, the patient should concentrate on imagining a specific movement instructed by a therapist, such Stroke is one of the most prevalent pathologies around the world. as wrist dorsiflexion. During MI, the contralateral motor cortex Stroke can cause devastating effects in survivors, including severe will exhibit event-related desynchronization (ERD), which is a motor and sensory impairments that hinder their activities of decrease of EEG bandpower in the µ (8–13 Hz) and β (16–30 Hz) daily living (Kim et al., 2020). The clinical consequences after a range. After the patient finishes performing MI, the contralateral stroke vary, depending largely on the location and the cause of motor cortex exhibits an increase in µ and lower β rhythm the damage (Prabhakaran et al., 2008). Diagnostic imaging tools activity, called event-related synchronization (ERS). An ERS can like Computational Tomography (CT) or Functional magnetic also occur during MI in the ipsilateral hemisphere in the µ range, resonance imaging (fMRI) are normally used to evaluate brain and is related to an idling state of those areas (Pfurtscheller damage in the acute and sub-acute phases, offering valuable and Aranibar, 1979; Graimann et al., 2002; Neuper et al., 2006; information about the diagnostic and functional prognosis for Kaiser et al., 2012). Many people with stroke exhibit atypical each case. Recent studies explored new methods to process and ERD/ERS activation patterns; for example, the affected cortex analyze brain signals acquired by conventional techniques like may be less excitable, and the changes in EEG activity may be electroencephalography (EEG) (Kanda et al., 2009; Leon-Carrion more prominent over nearby cortical areas (Neuper et al., 2006; et al., 2009; Foreman and Claassen, 2012; Rabiller et al., 2015; Wu Kaiser et al., 2012). et al., 2016) or magnetoencephalography (MEG) (Mäkelä et al., Hence, stroke patients often have abnormal changes in 2015; Ikkai et al., 2016; Krauss et al., 2018). ERD/ERS patterns resulting from MI. Kaiser et al. (2012) Quantitative EEG (qEEG) is a useful tool to extract features investigated how these abnormal patterns relate to the patient’s from the EEG signals and thereby help clinicians understand each functional state and spasticity, using a new parameter, the patient’s clinical state. qEEG parameters have shown multiple Laterality Coefficient (LC). For physical assessment, they used correlations with different pathologies, making qEEG an essential the European Stroke Scale (ESS), the Medical Research Council tool for different clinical fields (Nishida et al., 2011; Wang et al., (MRC) and the Modified Ashworth Scale (MAS). The LC 2013; Faust et al., 2015; Jeong et al., 2016; Muniz et al., 2016; presented significant correlations with the MRC scale and MAS. Piano et al., 2017). One qEEG parameter is the Brain Symmetry The findings of Kaiser et al. (2012) showed that high percentage Index (BSI), described by van Putten et al. (2004) to assess the changes in ERD patterns in the contralesional hemisphere are stroke risk during carotid endarterectomy surgery in real time related to a high degree of impairment. However, the aim of (van Putten et al., 2004; van Putten, 2006, 2007). Subsequently, the current publication is to evaluate a novel analysis studying Agius Anastasi et al. (2017) used the BSI with stroke patients and the correlation of LC with other functional scales like the Fugl- found correlations between this parameter and functional scales. Meyer assessment. The idea of the BSI is to assess the symmetry between both brain Here, we explore two different qEEG parameters and their hemispheres by using the EEG. relationship with the diagnosis and functional prognosis of stroke EEG can measure brain signals with a high temporal patients. One group of healthy participants and one group resolution, allowing clinicians to monitor brain activity in real of stroke patients participated in the study. Stroke patients time (Dobkin, 2007; McFarland and Wolpaw, 2017). Brain signals performed functional assessment sessions, and BCI rehabilitation can be read with a software program to provide the user with therapy for the upper extremity. EEG was recorded in two an external pathway for these brain outputs (Wolpaw, 2007). different situations: 8 min of resting state with open eyes This approach has been employed in numerous Brain Computer (rEEG), and MI using a BCI system for motor rehabilitation. BSI Interface (BCI) systems providing real-time communication and perameters were analyzed with rEEG, whereas LC was calculated control. BCIs have been used to control devices such as a during the MI period. To assess each patient’s functionality wheelchair (Carlson and del, 2013), prosthesis or functional before and after the therapy, we primarily used the Fugl- electrical stimulator (FES) (Ramos-Murguialday et al., 2013), Meyer assessment (FMA) (Gladstone et al., 2002; Woytowicz sometimes in combination with immersive feedback relating to et al., 2017). We also used eight other standardized tests used rehabilitation (Shokur et al., 2018). Over the past several years, in rehabilitation to assess motor function, spasticity, cognitive many publications have combined BCI, FES and other feedback function, and other parameters: Fahn Tremor Rating Scale, MAS, devices to increase cortical plasticity in stroke survivors helping Barthel Index, Box and Block Test, 9 Hole Peg test, 2 Point them regain movement control (Dobkin, 2007). Discrimination Test, Montreal Cognitive Assessment and Self- In this approach to movement restoration, stroke survivors rated questionnaire. This is the first study to employ such a broad perform Motor Imagery (MI) exercises during EEG recording range of tests along with analyses of BSI and other EEG-based (Cervera et al., 2018). The decoded brain oscillations can be parameters across several therapy sessions. used to move a virtual reality avatar or trigger an FES device to reproduce the imagined movement with the paretic limb (e.g., Cho et al., 2016; Irimia et al., 2017). These types of MATERIALS AND METHODS rewarding feedback only occur if the patient imagines the desired movement, providing a closed-loop feedback system for patients Participants and an objective means to monitor patient compliance for Thirty-two healthy subjects and thirty-six stroke patients with therapists and scientists. upper extremity hemiparesis were recruited for this study. Frontiers in Neuroscience | www.frontiersin.org 2 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation All healthy participants were volunteers recruited through the feedback presented on the monitor. Each session contained up Universitat de Vic, Spain. The stroke patients were recruited in to three runs of 80 trials each, depending of the patient’s fatigue. the rehabilitation center RecoveriX Gym in Schiedlberg, Austria. At the end of each session, the cap and FES pads were removed, Two patients dropped out from the study because of personal and the skin was cleaned with a moist cloth. Each session required problems that prevented them from attending recording sessions. about 60 min total, including preparation and cleaning. The patients’ characteristics are reported in the results section. The inclusion criteria for stroke patients were: (i) residual BCI System Description hemiparesis, (ii) the stroke occurred at least 4 days before the first The BCI system used in this study was RecoveriX (g.tec medical assessment, (iii) functional restriction in the upper extremities. engineering GmbH, Austria) (Irimia et al., 2016). This system Additionally, for all participants, the following criteria were managed all EEG data recording and real-time interactions with applied: (iv) able to understand written and spoken instructions, the patient and therapist, including visual feedback using a virtual (v) stable neurological status, (vi) willing to participate in the reality avatar and proprioceptive feedback using FES. Participants study and to understand and sign the informed consent, (vii) wore EEG caps with 16 active electrodes at positions FC5, FC1, able to attend recording sessions, (viii) no cerebellar lesions, FCZ, FC2, FC6, C5 C3, C1, CZ, C2, C4, C6, CP5, CP1, CP2, and (ix) no Botulinum toxin treatment for spasticity during the CP6, according to the international 10/10 system (extended 10/20 study. The participants were not recruited based on the scores system). A reference electrode was placed on the right earlobe and of any functional assessment. Ethics approval was obtained from a ground electrode at FPZ. the Ethikkommission des Landes Oberösterreich in Austria for Two FES electrodes were placed on the skin over the wrist the patients (#D-42-17), and the ethics committee of Comitè extensors of the left and right forearms. The frequency was set d’Ètica de la Recerca-CER of Universitat de Vic (Spain) for the to 50 Hz, and the pulse-width was set to 300 µs. Then, the healthy controls. stimulation parameters were individually adjusted for session until either (1) the optimal passive movement without pain for Protocol patients with mild or moderate muscle spasm, or (2) muscle Healthy Controls contraction was observed in the target muscle of the paretic side The healthy controls sat in a comfortable chair for 8 min for patients with severe muscle spasm. while rEEG was collected. During this resting state assessment, MI Exercise participants were asked to avoid unnecessary movements and Figure 1C depicts the timing of each trial. Each trial starts with keep their eyes open, aside from normal blinking. a beep, to help the participant focus on the upcoming task. Two seconds later, the system presents the instruction. For 1.5 s, the Stroke Patients system presents an arrow pointing to the left or right on the Each stroke patient participated in four assessment sessions and patient’s monitor and the word “left” or “right” in the participant’s 25 therapy sessions. mother tongue via headphones. These simultaneous visual and Assessment: A clinician assessed each patient twice before auditory cues direct the patient to imagine dorsiflexion of the the therapy began and twice after the last therapy session. Each left or right wrist (in pseudorandom order). The participant of these four assessment sessions had two components: (1) is instructed to start the MI immediately after receiving the the clinician recorded 8 min of rEEG with the same settings command and to continue the MI until the relax command is as described above for healthy controls and (2) the clinician presented auditorily. The feedback phase starts at the same time tested the patient’s motor function. The Pre1 assessment was as the instruction ends, that is 3.5 s after trial begin. The feedback performed 1 month before starting the therapy, and the Pre2 devices can only be activated during this phase. assessment was performed just before the therapy started. The Post1 assessment was performed just after the last session, and Feedback Presentation the Post2 assessment occurred 1 month after the last session. One Visual and proprioceptive FES feedback are provided in hundred and and thirty six assessment sessions were performed synchrony and only in the feedback phase. Classification of motor in total (4 per patient). imagery is done every second. If the classified MI matches the Therapy: Patients completed 25 sessions, with two presented command (left or right), then feedback is switched sessions per week. on, which means the wrist dorsiflexion is initiated by electrical Figures 1A,B depict different system components and the stimulation and presented visually on the computer screen. physical layout during each therapy session. At the beginning During incorrect classifications, the initiated movement is done of each therapy session, the therapist talked with each patient in the opposite direction. For classification, we used linear to confirm that the patient understands the MI task and the discriminant analysis (LDA) on the spatially filtered data. We upcoming procedure. Next, the EEG cap and FES pads were followed the steps described in Irimia et al. (2018) except for a placed on the patient, and FES parameters were adjusted, as change in the electrode setup: FPZ, FP1, FP2, AF7, AF3, AFZ, detailed below. After this preparation, the patient was seated in AF4, AF8, F7, F5, F3, F1, FZ, F2, F4, F6, F8, FT7, FC5, FC3, a comfortable chair in front of a table, facing a monitor where FC1, FCZ, FC2, FC4, FC6, FT8, T7, C5, C3, C1, CZ, C2, C4, two virtual arms were projected in a position and orientation C6, T8, TP7, CP5, CP3, CP1, CPZ, CP2, CP4, CP6, TP8, P7, mimicking the subject’s arms. The patient was asked to place both P5, P3, P1, PZ, P2, P4, P6, P8, PO7, PO3, POZ, PO4, PO8, O1, hands on the table and perform MI while following cues and OZ, O2, O9, O10. Frontiers in Neuroscience | www.frontiersin.org 3 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation FIGURE 1 | BCI system components. (A) Shows the motor learning loop. (B) System setup. (C) Trial description. Event-Related Synchronization and produces a desynchronization (event-related desynchronization Desynchronization or ERD) of cortical motor neurons, showing a decrease in the Figure 2 presents ERD/ERS patterns that typically occur during bandpower of the waves with a frequency of 8–13 Hz (mu MI. This figure was generated using the data from one BCI frequency rhythm). The ipsilateral motor cortex shows ERS training session. During MI, the contralateral motor cortex patterns to suppress corresponding motor areas during MI of the Frontiers in Neuroscience | www.frontiersin.org 4 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation opposite hand side (Pfurtscheller and Aranibar, 1979; Graimann deviation for each channel was rejected from the BSI calculation. et al., 2002; Neuper et al., 2006; Kaiser et al., 2012). To create such We did consider an algorithm for rejection of EOG related maps, the change of EEG bandpower of several bandpass filtered artifact, but determined that it would be unnecessary based on frequency bands was calculated and plotted. The frequency bands analysis of pilot data. The most frontal EEG electrodes are in the chosen here ranged from 8 to 30 Hz in steps of 2 Hz. In row of FCZ; hence, the influence of EOG was small. Furthermore, each band, the power was calculated stepwise in windows of 16 participants were asked to focus on the screen during the samples (0.0625 s). Then, the bandpower of each window was reference period, thus limiting eye movement. Movement related compared to the bandpower of the reference period (gray area artifacts were found and rejected with our approach based on the in Figure 2), during which the participant was in a resting state. standard deviation. Finally, the Fourier coefficient was calculated The comparison used the following formula, in which A is the from the power density estimation using the Welch method. bandpower of one single window and R the bandpower within the reference period: Laterality Coefficient A−R The raw EEG data recorded during the MI sessions was used to ERD = ∗ 100% calculate the LC parameter. The LC coefficient was calculated for R each session twice: first for trials of MI of the paretic (p) hand Finally, a bootstrapping significance test (α = 0.05) was done and again for trials of the healthy (h) hand. We employed the for all windows. Values that are not significant were set to following formula, where C and I refer to the contralateral and 0 and are plotted in white in Figure 2. High ERD values ipsilateral values of the ERD/ERS patterns during the MI. (decreased bandpower) are plotted in red, whereas high ERS values (increased bandpower) are plotted in blue. LCp/h = (C − I)/(C + I) EEG Parameters We followed six steps to calculate C and I: Brain Symmetry Index The BSI is a parameter that compares the spectral power of (1) Band filtering (8–13 Hz or 13–30 Hz) of the EEG signal; the two hemispheres of the brain using bandpass filtered EEG (2) Frame artifact rejection if a sample overflows a threshold signals. The BSI value ranges from 0 to 1 and is a measure of the based on the median variance among the samples of all symmetry between both hemispheres. A BSI value of 0 reflects the frames; total symmetry and 1 total asymmetry. The BSI value should be (3) Laplacian derivation using the surrounding electrodes; closer to 0 in healthy people and higher in stroke patients. (4) ERD/ERS patterns calculation according to (Graimann et al., 2002); Method (5) Summation of all ERD/ERS values from second 2 until the The BSI of a segment of EEG is calculated using a revised BSI end of the ERD map (second 8); and formula (van Putten, 2007), which is based on the squared value (6) Apply the formula to obtain the LC coefficients. of the Fourier coefficients: K 1 X R∗n (t) − L∗n (t) Assessment Tests BSI(t) = K R∗n (t) + L∗n (t) We used 10 tests to assess each patient’s functional capabilities n=1 during each of the four assessment sessions. All of these with tests are well-established in the scientific literature and clinical M 1 X 2 practice. These ten tests each measure different aspects of R∗n (t) = αn (ch, t) M motor function, other motor impairment (tremor and spasticity), ch=1 cognitive function, sensory discrimination, and self-reported where α2n (ch, t) is the Fourier coefficient with index n of channel impact. The Supplementary Material provides a complete ch at time t. For the right hemisphere R∗n (t), the same formula is description of each test. applied for the left hemisphere electrodes. The scales used for the motor assessment were: Fugl-Meyer We collected resting state data from 16 EEG sites. For the BSI Assessment (FMA) for the upper extremity (FMAue) and for the calculation, we discarded the central sites and split the remaining lower extremity (FMAle), Box and Block Test (BBT) and 9 Hole sites in two sets: right and left. For the left hemisphere, the Peg Test (HPT) for the healthy and paretic hand. Each patient electrodes were: FC3, C5, C3, C1, CP3, and CP1. For the right completed the BBT and 9HPT tests with both hands to provide hemisphere, the electrodes were: FC4, C2, C4, C6, CP2, and CP4. an individualized baselne. We processed 8 min of resting state EEG for each participant The scale used to assess the tremor was the Fahn Tremor (in the healthy and stroke groups). We bandpass filtered (1– Rating Scale (FTRS). We assessed tactile discrimination with the 25 Hz) the whole EEG, and then we cut it in frames of 4 s, with Two Point Discrimination Test (TPDT). The spasticity of the a 2-s overlap. We used a Hamming window to prevent spectral wrists and fingers was assessed with the Modified Ashworth Scale distortion. We used an artifact detection method based on the (MAS). The Montreal Cognitive Assessment (MOCA) was used overflow of the EEG standard deviation on each 4 s window to assess the cognitive status of the patients. The Barthel Index frame. Any frame with more than 1.5 times of the total standard (BI) was used to assess performance in daily life activities. Frontiers in Neuroscience | www.frontiersin.org 5 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation FIGURE 2 | ERD maps obtained after one session of BCI training. Top right: ERD map during right hand MI on site C3. Top left: ERD map during right hand MI on C4. Bottom right: ERD map during left hand MI on C3. Bottom left: ERD map during left hand MI on C4. Each plot shows the time from 0 to 8 s (x-axis) and frequencies between 8 and 30 Hz (y-axis). Red areas indicate high ERD. Blue areas mark the opposite: an ERS. Vertical bars indicate the cue onset at 2 s. Patients also completed a self-rated questionnaire (SRQ) to t-test or the Wilcoxon test was used, depending on the assess pain, function, memory, thinking, mobility and the home normality assumption. and community, and stroke recovery. For comparisons between multiple groups, Welch’s ANOVA was used. We used Welch’s ANOVA for comparisons between multiple groups. When Welch’s ANOVA yielded significant results, we conducted a post hoc analysis using the Single-step Statistical Analysis Games-Howell test. The sample size selection was based on the previous literature A correction for multiplicity (i.e., multiple hypotheses testing) (Kaiser et al., 2012). The statistical analyses were performed using was not utilized, because no final conclusion and decision on the MATLAB R2017a (MathWorks Inc., United States). Normality correlation of qEEG parameters and functional scales is drawn. of data was tested using the Shapiro–Wilk test. The statistical Similarly, while a statistical analysis concerning the functional test was chosen according to the normality of the sample, the scales before and after BCI therapy was carried out, the goal homogeneous of variance (Levene’s or Brown–Forsythe test of of this paper is not to establish definitive proof on the efficacy equal variance) and sample size. Leven’s test of equal variance of BCI-based therapy. Concordantly, we did not correct for was used if normality could be assumed; otherwise, the Brown– multiplicity in this statistical analysis either (Bender and Lange, Forsythe test of equal variance was used. Descriptive statistics are 2001). Additionally, Hurlbert and Lombardi (2012) recommend reported as mean and the standard deviation (SD), or median and carrying out statistical tests without any adjustments for multiple the inter-quartile range (IQR) of 0.25 and 0.75. hypothesis. This recommendation stems from the reasoning that The correlation tests were chosen according to the results the true probability of type I errors occurring is very small, obtained by the Shapiro–Wilk test. When the normality because most null hypotheses can be expected to be false based assumption was not rejected, the correlation was made using on prior knowledge (Hurlbert and Lombardi, 2012). In the Pearson’s linear correlation test; otherwise, the correlation was present case, one can reasonably accept this premise, based on done using Spearman’s rank correlation test. the available literature discussed in the section “Introduction.” For the two group comparisons (for example, BSI comparison based on gender), the test was selected based on the normality and homogeneity of variance of the samples and the independence RESULTS assumption. The unpaired t-test was used for comparisons of two independent likely normally distributed groups. For Participants’ Baselines the comparisons before-and after the treatment [Pre2 vs. Thirty-two healthy subjects were enrolled in the study, with 13 Post1] with measures from the same population, the paired males and 19 females. The mean age in the healthy group was Frontiers in Neuroscience | www.frontiersin.org 6 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation TABLE 1 | Participants’ baselines. TABLE 2 | Results of BSI-based age analysis. Group n Age (y) SD Male Female One-way ANOVA Healthy 32 42.3 15.4 13 19 SS Df MS F P Patient 34 65.3 14.4 22 12 Groups 0.000983 2 0.000492 0.3843 0.6844 Cortical 5 57.6 27.3 4 1 Error 0.0371 29 0.0013 Subcortical 17 66.4 12.7 9 8 Total 0.0381 31 Cortical + Subcortical 12 67.0 09.4 9 3 according to the Shapiro–Wilk test (P = 0.117). We explored 42.3 years (SD = 15.4). Thirty-four stroke patients participated the relationship between BSI and age using Pearson’s method (excluding two who dropped out). Twenty-two of the patients and One-way ANOVA. Figure 3A shows that the Pearson’s were male (64.7%), and the other 12 stroke participants were correlation did not show significant correlation between BSI and female (35.3%). The stroke patients’ mean age was 65.3 years age (ρ = −0.110, P = 0.548). Subsequently, we compared BSI (SD = 14.4); this difference in age will be addressed at a later stage across age groups (under 30 years, between 30 and 50 years and of this analysis. over 50 years). The variance of each group, using Levene’s test, did Table 1 shows the participants’ baselines. The stroke not show significant results (df = 29.00, F = 1.338, P = 0.278). The participants were classified in three groups based on their stroke analysis of variance shows that there is no significant difference diagnosis: Cortical, Subcortical, Cortical + Subcortical. The most in the BSI parameter based on the three age groups (F = 0.3843, common type of stroke was Subcortical with 17 patients (50.0%), P = 0.684). See Figure 3B and Table 2. followed by Cortical+Subcortical with 12 patients (35.3%) and Cortical with 5 patients (14.7%). Twenty-seven of these patients BSI Based on Gender were in the chronic phase (79.4%), and only 7 in the subacute Figure 4 presents the results of this subgroup analysis. Both phase (20.6%). Twenty three patients had a stroke in the right groups have similar variance (Levene’s test results: df = 30.00, hemisphere (67.7%), and the stroke was in the left hemisphere in F = 1.733, P = 0.198). The result of this analysis shows that there 11 patients (32.4%). is a statistical difference in BSI based on gender, according to the unpaired t-test (t-value = | 2.333|, P = 0.027). Brain Symmetry Index (BSI) BSI Differences Between Age Groups BSI Between Groups To date, there is no evidence to demonstrate the variability of BSI Since the results obtained in the BSI based on age did not show with age. We performed a statistical analysis using the rEEG data significant differences in the healthy group, we compared the BSI from the healthy subjects. The data follows a normal distribution between groups (stroke and healthy) despite the age difference. FIGURE 3 | BSI analyses based on age. (A) Shows the correlation between age and BSI using Pearson’s method. The yellow line is the regression line and the gray lines are the CI at 95%. (B) Shows the BSI results based on three clusters. Frontiers in Neuroscience | www.frontiersin.org 7 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation FIGURE 4 | BSI subgroup analyses based on gender in the healthy group. Mean and SD of each group: Male = 0.1272 (SD = 0.0278); Female = 0.0997 (SD = 0.0357). Significant difference between groups using unpaired t-test; t-value = | 2.333| P = 0.027. We first analyzed the resting state data collected during the summarizes results from each group. Welch’s ANOVA test found assessments, consisting of 136 assessment sessions from 34 stroke an associated probability of P = 0.003 and F = 8.929, so the patients and 32 EEG recordings from 32 healthy subjects. We hypothesis of equal sample means was rejected. As the Welch’s calculated the BSI of each assessment (Pre1, Pre2, Post1, Post2) ANOVA test showed significant results, the Games-Howell test from each patient’s rEEG data. For this analysis, we used the was used to complete the analysis. The single-step Games-Howell median of Pre2 and Post1. In one case, the BSI could not test (see Table 4) shows significant differences between group be calculated due the corruption of the EEG data recordings 1 (healthy group) and both group 3 (subcortex group) and (this patient belongs to Cortical group). The first step was the group 4 (cortex + subcortex group). Figure 5 shows the BSI normality testing of each group. The first three datasets did not values for each group. attain significance using the Shapiro–Wilk Test, hence, a normal distribution can be assumed. Group 4 (cortex + subcortex) is Correlations Between BSI and Functional Tests the only one that is not normally distributed. The equality of the Figure 6 shows a significant correlation between BSI and patients’ variances cannot be assumed (Levene’s Test results: df = 61.00, outcomes on the FMAue scale. The correlation coefficient of this F = 5.798, P = 0.001). We used Welch’s ANOVA test to compare the BSI parameter across the four groups, because this method is reasonably robust to deviations of normality, when the variances are substantially different and even if the sample sizes are unequal TABLE 4 | Single-step Games-Howell test. (Kohr and Games, 1974; Tomarken and Serlin, 1986). Table 3 Comparison Delta SE Df P H lb ub 1 2 −0.068 0.045 4.157 0.505 0 −0.246 0.110 TABLE 3 | BSI analysis summary statistic. 1 3 −0.047 0.010 33.870 0.001 1 −0.075 −0.019 1 4 −0.082 0.023 12.708 0.019 1 −0.151 −0.013 Group Size Mean Variance 2 3 0.021 0.045 4.275 0.965 0 −0.156 0.198 2 4 −0.014 0.050 6.171 0.992 0 −0.185 0.156 1 32 0.1109 0.0012 3 4 −0.035 0.024 13.937 0.481 0 −0.105 0.035 2 4 0.1789 0.0098 3 17 0.1580 0.0011 Result of group comparison based on BSI values using Games-Howell test. The 4 12 0.1931 0.0061 first column shows the group code; 1 – Healthy group, 2 – Cortical group, 3 – Subcortical group, 4 – Cortex + Subcortex group. The column ‘H’ shows the Group numbers: 1 – Healthy, 2 – Cortical, 3 – Subcortical, 4 – significant (H = 1) and non-significant (H = 0) differences at alpha level, and the Cortical + Subcortical. column ‘P’ shows the significance level of these comparisons. Frontiers in Neuroscience | www.frontiersin.org 8 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation FIGURE 5 | BSI values in resting state with open eyes for each group. (*) Indicates the significant differences based on the Games-Howell test. The BSI of the healthy group is significantly different from the subcortical group and cortex + subcortex group. Healthy group median = 0.1173, IQR = 0.0799 – 0.1432. Cortical group = 0.1739, IQR = 0.1043–0.2534. Subcortical group = 0.1507, IQR = 0.1349–0.1890. Cortex + Subcortex group = 0.1612, IQR = 0.1400–0.2110. relationship is −0.430 and P = 0.046. Lower BSI values are related alpha (LChα and LCpα) and beta (LChβ and LCpβ) bands. We to better functionality. explored the LC (α and β) between groups (Cortical, Subcortical, and Cortical + Subcortical), and found no significant differences Laterality Coefficient of LC between groups (Welch ANOVA, F = 0.36, P = 0.7033). The LC was calculated separately for the MI tasks of the healthy In this part of the analysis, we correlated the LC mean of the (LCh) and paretic (LCp) hand. We calculated the LC for the 25 BCI therapy sessions against the mean of the results from motor tests collected in the Pre2 and Post1 assessment visits. The Shapiro–Wilk Test shows that the data are likely not normally distributed at alpha level. The Spearman test has been used for the correlation analysis. Table 5 shows the correlation’s results of LC against the functional scales. Alpha Band The LC calculated during the MI task with the healthy hand (LChα) is the parameter that shows the highest correlation with functional scales. In general terms, the results show that LC values near 0 are related to better functionality and less tremor in the paretic upper extremity (see Figure 7). Tremor of the paretic hand assessed by FTRS shows a significant correlation with the LChα. The correlation coefficient is positive (ρ = 0.450 and P = 0.008). Thus, low degrees of tremor are related to LChα values near to 0. In the BBT of the paretic hand, there is a stable correlation with all the LC parameters and bands. Here, the LChα shows a correlation but with a negative sign. The correlation is strong (ρ = −0.616 and P < 0.001). This correlation shows that good FIGURE 6 | Correlation between BSI and FMA upper extremity. The gray lines scores in the grasp ability, as assessed by BBT, are related to low are the confidence interval at 95%. values of LChα. Frontiers in Neuroscience | www.frontiersin.org 9 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation TABLE 5 | Significant correlations between LC and functional scales using Spearman Correlation are colored red. Scale Laterality Coefficient α β LCh LCp LCh LCp Name Side ρ P ρ P ρ P ρ P BI − −0.260 0.138 0.058 0.743 −0.154 0.383 0.184 0.296 FTRS Healthy −0.038 0.829 0.105 0.555 −0.093 0.600 0.060 0.734 Paretic 0.450 0.008 −0.245 0.162 0.336 0.052 −0.490 0.003 MAS Wrist 0.076 0.670 −0.216 0.220 −0.116 0.514 0.034 0.848 Fingers 0.237 0.176 −0.262 0.134 0.109 0.539 −0.099 0.579 BBT Healthy 0.102 0.566 −0.154 0.386 0.059 0.741 −0.141 0.425 Paretic −0.616 <0.001 0.354 0.043 −0.418 0.016 0.569 0.001 9HPT Healthy −0.042 0.813 0.036 0.839 −0.167 0.345 0.186 0.291 Paretic 0.536 0.236 −0.357 0.444 0.714 0.088 −0.607 0.167 TPDT Thumb H −0.169 0.340 0.031 0.862 0.038 0.830 −0.005 0.979 Index H −0.010 0.956 0.041 0.820 −0.053 0.765 0.157 0.374 Thumb P 0.000 0.999 −0.067 0.746 −0.139 0.499 0.152 0.459 Index P 0.065 0.751 −0.079 0.701 −0.082 0.689 −0.125 0.543 FMAue – −0.706 <0.001 0.400 0.019 −0.440 0.009 0.384 0.025 FMAle – −0.601 0.006 0.271 0.261 −0.252 0.298 −0.057 0.817 SRQ Pain 0.287 0.100 −0.157 0.374 0.095 0.591 −0.115 0.518 Function −0.427 0.012 0.316 0.069 −0.488 0.003 0.447 0.008 Memory −0.068 0.704 −0.130 0.465 −0.226 0.198 −0.033 0.855 Mobility −0.216 0.219 −0.034 0.849 −0.150 0.396 0.033 0.855 Recovery 0.065 0.717 −0.205 0.245 0.083 0.642 0.061 0.732 MOCA – 0.005 0.982 −0.032 0.884 −0.232 0.288 0.043 0.847 Also, the LChα parameter showed significant correlations with low values of LChβ (ρ = −0.418 and P = 0.016) and values near the FMA upper and lower extremity. The FMAue correlation has 0 in LCpβ (ρ = 0.569 and P = 0.001). The general motor function a stronger correlation coefficient (ρ = −0.706 and P < 0.001) than of the upper extremity, assessed by FMA, is also correlated with the FMAle (ρ = −0.601 and P = 0.006). The correlation coefficient LChβ (ρ = −0.440 and P = 0.009), and with LCpβ (ρ = 0.384 is negative in both cases, and these results are consistent with the and P = 0.025). other relationships explained above – better motor function in Finally, the last significant correlation is between the function the lower and upper extremity, as assessed by FMA, is related to scale part of SRQ and LChβ (ρ = −0.488 and P = 0.003) and LCpβ LChα values near to 0. (ρ = 0.447 and P = 0.008). Again, the best functionality is related Finally, the LChα is also correlated with the function score of to values near to 0 of both LC parameters. All correlation results the SRQ (ρ = −0.427 and P = 0.0212). The function score of SRQ regarding the LC and functional scales can be seen in Table 5. is based on the subjective opinion of the patient doing different motor tasks. The negative correlation shows that good scores in Assessment Tests Before and After BCI the function score of SRQ are related to low values of LChα. Treatment Other similar correlations with opposite signs have been found The results in this section summarize differences from the Pre2 for the LCpα. In this case, LCp values near to 0 are related to to Post1 assessments across different tests. We used the Wilcoxon better performance in the FMAue score (ρ = 0.400 and P = 0.019) signed rank test for statistical analysis, since the data did not and also in the BBT of the paretic hand (ρ = 0.354 and P = 0.043). present a normal distribution (see Table 6). The improvement of each scale is presented using the median and IQR, and the mean Beta Band and SD are also provided if differences are significant. LChβ and LCpβ also presented some interesting correlations with the functional scales. In general, the correlations found in this Fugl-Meyer Assessment frequency band are weaker than the correlations found in the The FMAue test has a score range of 0–66. One of the 34 alpha band. The low tremor degree in the paretic hand assessed patients had only slight hemiparesis and attained the maximum by FTRS (higher scores in this scale) is correlated with values near FMAue score in the pre-assessment. The Wilcoxon signed rank to 0 in LCpβ (ρ = −0.490 and P = 0.003). The good grasp ability in test shows that there is a significant improvement in FMAue the paretic hand, assessed by BBT (BBT_p), is also correlated with after the therapy (1FMAue = 1 [0–8], P = 0.002). The mean Frontiers in Neuroscience | www.frontiersin.org 10 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation FIGURE 7 | Significant correlations between LC of the paretic hand in alpha band and functional scales. The gray lines are the confidence interval at 95%. (A) Correlation between LChα and BBT of the paretic hand. (B) Correlation between LChα and FTRS of the paretic hand. (C) Correlation between LChα and FMAue of the motor part. (D) Correlation between LChα and FMAle of the motor part. improvement is 3.12 (SD = 5.1). Twenty one patients (61.8%) degree of tremor in the paretic hand (FTRS_p) before the therapy. improved at least 1 point in the FMA score. Among patients who After the therapy, 10 of these 32 patients (31.3%) exhibited a improved, the mean improvement was 5.76 points (SD = 4.61). decreased tremor in the paretic hand. One of these 32 patients Six patients (17.7%) decreased at least 1 point in FMA score, and (3.1%) showed an increase of tremor after the therapy. The other the mean decrease in this group was −2.5 points (SD = 1.76). The patients did not report any changes. remaining 6 patients (17.7%) had an improvement equal to 0. Modified Ashworth Scale Barthel Index The MAS scale showed a statistical reduction of the spasticity in The BI did not show significant improvements after the therapy the wrist [1MAS_w = 0 [-1-0], P = 0.003, mean improvement (1BI = 0 [0–5], P = 0.480). The BI score decreased in 7 patients −0.37 (SD = 0.69)], and in the fingers [1MAS_f = 0 [-1-0], (20.6%), 11 patients (32.4%) reported positive changes in the BI P = 0.001, mean improvement −0.41 (SD = 0.63)]. Twenty three after the therapy, and 16 patients (47.1%) did not show changes of the 34 patients (67.7%) reported some spasticity in the wrist in this parameter. (MAS > 0), and 25 patients (73.5%) reported some spasticity in the fingers. Twelve of the 23 patients (52.2%) who reported Fahn Tremor Rating Scale wrist spasticity prior to therapy reported a decrease after therapy. The FTRS for the healthy hand (FTRS_h) did not show Fourteen of the 25 patients (56.0%) who reported finger spasticity a significant difference before versus after the BCI therapy prior to therapy reported a decrease after therapy. (1FTRS_h = 0 [0–0], P = 0.984). The FTRS in the paretic hand (FTRS_p) did show a significant improvement (1FTRS_p = 0 Box and Block Test [−1-0], P = 0.018). The mean improvement of FTRS_p is −0.65 We observed a statistical improvement of BBT in the healthy (SD = 1.5). Thirty two of the 34 patients (94.1%) reported some hand [1BBT_h = 2 [1–8] and P = 0.005, mean improvement 3.64 Frontiers in Neuroscience | www.frontiersin.org 11 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation TABLE 6 | Changes in the functional scales. Scale SWT Pre Post 1 P H P Median [IQR] Median [IQR] Median [IQR] Mean (SD) BI 1 0.001 85 [70–95] 85 [65–100] 0 [0-5] 0.29 (6.15) 0.480 FTRS_h 1 <0.001 0 [0–0] 0 [0–0] 0 [0-0] −0.03 (0.83) 0.984 FTRS_p 1 <0.001 12 [7–12] 12 [4–12] 0 [-1-0] −0.65 (1.52) 0.018 MAS_w 1 <0.001 1.25 [0–3] 1 [0–2] 0 [-1-0] −0.37 (0.69) 0.003 MAS_f 1 0.001 2 [0–3] 1 [0–2] 0 [-1-0] −0.41 (0.63) 0.001 BBT_h 0 0.346 54 [44–68.75] 54.5 [47–73] 2 [1-8] 3.64 (7.4) 0.005 BBT_p 1 <0.001 0 [0–3.5] 0 [0–4.25] 0 [0-1] 1.22 (3.5) 0.034 9HPT_h 1 <0.001 23 [18.75–27.25] 23 [20–26] −1 [-2-2] −1.06 (4.6) 0.325 9HPT_p 1 0.046 164 [76–346.25] 135.5 [93–324] −24.5 [-70-2] −34 (45.35) 0.375 TPDT_h_t 1 <0.001 3 [2–4] 3 [2–4] 0 [-0.75-1] 0.06 (1.34) 0.720 TPDT_h_i 1 0.001 3 [3–4] 3 [3–4] 0 [-1-0.5] −0.09 (1) 0.667 TPDT_p_t 0 0.127 4 [3–5] 4 [3–5] 0 [-1.25-1] −0.08 (1.75) 0.888 TPDT_p_i 1 0.001 3 [3–4.5] 4 [3–4.75] 1 [-0.25-1] 0.46 (1.66) 0.324 FMAue_m 1 0.003 19 [10–37] 23 [12–41] 1 [0-8] 3.12 (5.06) 0.002 FMAle_m 0 0.443 17 [9.75–24.25] 19 [9–25] −0.5 [-2-2.5] 0.56 (4.1) 0.856 Pain 0 0.275 25 [16–37] 21 [15.5–36.5] −2 [-7.5-6] −3.75 (11.93) 0.285 Function 1 <0.001 3 [0–12] 6 [0–12] 0 [0-4] −3.46 (16.71) 0.444 Memory 1 0.003 55 [39.75–70] 59 [43.5–68] 0 [-3-6.5] 1.63 (12.89) 0.614 Mobility 1 0.023 67 [35–79] 70.5 [43.5–81.5] 5 [-3-10.5] 5.93 (17.54) 0.056 Recovery 0 0.174 5 [4–8] 6 [5–7] 0 [-2-3] 0.6 (2.91) 0.311 MOCA_Total 1 0.043 24.5 [17–27] 26 [21–27.75] 2.5 [0-6] 2.71 (3.29) 0.012 The first column shows the results of the Shapiro–Wilk Test (SWT), to assess the normality of the dependent variable. The last column (P) presents the probability results of the paired test, with statistically significant differences colored red. (SD = 7.4)]. The changes in the paretic hand are also significant in the healthy index finger. Six patients (17.7%) improved in [1BBT_p = 0 [0–1] and P = 0.034, mean improvement 1.22 the TPDT at least by 1 mm in the paretic thumb, and three (SD = 3.5)]. Ten patients (29.4%) improved the BBT score with patients (8.8%) reported at least 1 mm of improvement in the paretic hand, 2 patients (5.89%) decreased the BBT score the paretic index. with the paretic hand, and 22 patients (64.7%) did not change from the initial BBT score. In three cases (8.8%), it was impossible SRQ to perform the BBT before the therapy due to the severity of the Sixteen patients (47.1%) reported at least 1 point of pain motor impairment, but after the therapy, these patients could reduction, and 11 patients (32.4%) reported at least 1 point move at least 1 block in the BBT. of pain increase. Thirteen patients (38.2%) reported an improvement in the ability to perform ADLs, and six patients 9HPT (17.7%) reported a decrease in ADL performance. Eleven patients The 9HPT in the paretic hand is one of the most commonly used (32.4%) reported an improvement in the memory part, and 11 tests of grasp function. Only five patients (14.7%) could perform patients (32.4%) reported a decrease in memory. Sixteen patients the test before the therapy, and six patients (20.6%) could perform (47.1%) reported an improvement in the mobility part of the this test after the treatment. No significant improvements were questionnaire, while 11 patients (32.4%) reported a decrease in observed after the therapy in the healthy hand (19HPT_h = −1 mobility. Finally, 12 patients (35.3%) reported a better general [-2-2], P = 0.325), or in the paretic hand (19HPT_p = −24.5 [- recovery after BCI therapy, and 10 patients (29.4%) reported 70-2], P = 0.375). The results show that the time in the healthy a worse recovery after BCI therapy. There are no significant hand has slightly decreased, and in the affected hand the decrease changes in any part of SRQ (1Pain = −2 [-7.5-6], P = 0.285; in time was great. 1Function = 0 [0-4], P = 0.444; 1Memory = 0 [-3-6.5], P = 0.614; 1Mobility = 5 [-3-10.5], P = 0.056; 1Recovery = 0 Two Point Discrimination Test [-2-3], P = 0.311). This test did not show significant changes before vs. after the therapy in the thumb or index of the healthy hand MOCA (1TPDT_h_t = 0 [-0.75-1], P = 0.720; 1TPDT_h_i = 0 [-1- The comparison between before and after the therapy showed a 0.5], P = 0.667). The paretic hand did not show a significant significant improvement in cognitive function, 1MOCA = 2.5 [0- improvement either (1TPDT_p_t = 0 [-1.25-1], P = 0.888; 6], P = 0.012, with a mean improvement of 2.71 (SD = 3.29). Ten 1TPDT_p_i = 1 [-0.25-1], P = 0.324). Eight patients (23.5%) patients (29.4%) improved by at least one point after therapy, and improved the discrimination between two points in the healthy the MOCA score decreased in two patients (5.9%). The remaining thumb by at least 1 mm, and nine patients (26.5%) improved patients reported no change. Frontiers in Neuroscience | www.frontiersin.org 12 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation DISCUSSION LC in Alpha Band We calculated the LC using the event-related synchronization The objective of this experiment was to explore how two EEG- and desynchronization patterns generated during the MI task based parameters relate to different facets of stroke diagnosis (Kaiser et al., 2012). The LC is derived in a similar manner as the and functional prognosis during BCI-based stroke rehabilitation BSI, but the LC yields results from −1 to 1. We calculated the LC therapy. The BSI was derived from EEG data recorded during the in two frequency bands, 8–13 Hz (α band, mu frequency rhythm) assessment visits in the resting state, while the LC was based on and 13–30 Hz (β band) and found the most relevant results in EEG data recorded during MI exercises. the alpha band. In general, LC values calculated during the MI tasks with the healthy hand (LCh) were between 0 and 1, while LC values of the paretic hand MI tasks (LCp) were between −1 BSI, Age, and Gender and 0. The LCh in alpha band presented numerous significant The analysis of the BSI in healthy subjects based on age correlations with functional scales. We also observed most of suggests that this parameter does not change with age. However, these significant correlations with the LCp parameter, but with this issue needs to be further explored in larger studies. The the opposite sign. results show significant difference in BSI based on gender; The LC values for the healthy hand presented noteworthy males usually have higher BSI values than females. These results correlations with four dependent variables. LCh values near 0 could help our understanding of the BSI parameter in healthy were related with a higher BBT score in the paretic hand, which conditions, improve detection of pathological values correlated indicates better grasp function (ρ = −0.616 and P < 0.001). with different brain affectations that can help with diagnosis The LCh was also significantly correlated with tremor, assessed of stroke and other conditions, and support further research by FTRS. Participants with LCh values near 1 tended to have involving gender differences. a higher FTRS score (reflecting greater tremor) in the paretic hand (ρ = 0.450 and P = 0.008). Finally, the LCh parameter was significantly correlated with the FMAue and FMAle. LCh BSI and Stroke Diagnoses values closer to 0 reflect better motor functionality for the upper Stroke patients were divided into three different groups based on extremity (FMAue, ρ = −0.706 and P < 0.001) and for the lower stroke location; Cortical, Subcortical and Cortical + Subcortical. extremity (FMAle, ρ = −0.601 and P = 0.006). The Cortical group was the smallest group with only five patients The correlations between the LCp and the functional scales are and exhibited the highest BSI variability. Prior work found less common than the LCh. This could occur because the affected similar results, with an almost identical boxplot distribution but hemisphere does not present a normal activation pattern due the a smaller sample size (Agius Anastasi et al., 2017). Our results stroke, but the healthy hemisphere maintains the normal patterns show that healthy participants had significantly lower BSI values of desynchronization during the ipsilateral motor movements than stroke patients of the Subcortical group (P = 0.001) and the (originated in the affected hemisphere). The ERD/ERS patterns Cortex + Subcortex group (P = 0.019); see Table 4 and Figure 5. observed in the healthy hemisphere should be more stable than The high variability in the Cortical group may be due to small the ERD/ERS patterns observed in the affected side of the brain. size of this subgroup. Moreover, in these patients the location of the lesion is very peripheral, and most of the neural activity observable via EEG originates from the cortex; consequently, the aberrant neural activity is more apparent in these patients than in LC in Beta Band the ones with other stroke locations. The LC calculated in the β band showed similar correlations Hence, despite the high variability in the Cortical group, the (see Table 5). Interestingly, LCβ shows significant correlations BSI parameter did differ significantly between the healthy control with the scales where more mental concentration is required vs. stroke groups Subcortical and Cortex + Subcortex. Results (FTRS and BBT). In both scales, values near 0 in LCβ (healthy were consistent with prior work (Agius Anastasi et al., 2017). and paretic) are correlated with better grasp ability and less With further research, the BSI could become a tool to support tremor. Other studies showed correlations between the EEG stroke diagnosis, including stroke location and severity. activity in beta band and concentration (Janssen et al., 2017; Kiiski et al., 2020). BSI and Functional Impairment We also analyzed the correlations between the values of BSI and the patient’s functional state (Figure 6). The most noteworthy Clinical Improvements Before vs. After correlations observed showed that patients with lowest BSI BCI Therapy have better motor function in the upper extremities (FMAue). The main objective of the study was not to demonstrate the The correlation between BSI and FMAue was also observed efficacy of the BCI system in neurorehabilitation, nor to compare in prior studies; lower BSI values were correlated with higher BCI-based therapy to other forms of therapy. The relationship functionality in the upper extremity (Agius Anastasi et al., 2017). between BCI stroke therapy and functional outcomes has been Thus, the BSI could be a useful parameter to assess functional addressed in numerous studies (Ramos-Murguialday et al., 2013; impairment during stroke assessment and rehabilitation. Pichiorri et al., 2015; Remsik et al., 2016; Biasiucci et al., 2018; Frontiers in Neuroscience | www.frontiersin.org 13 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation Cho et al., 2018). However, we would like to add to the existing such as medications or non-invasive brain stimulation; measure literature by discussing our observations. long-term changes via longitudinal follow-up assessments; and FMAue was the primary measure of motor function in this compare the utility of these measures to other tools. study. When assessing the motor function of the upper extremity by FMAue, we found the most important significant clinical improvement (1FMAue = 1 [0-8] and P = 0.002). On average, DATA AVAILABILITY STATEMENT the stroke patients improved by 3.21 points (SD = 5.1) in the FMAue with the BCI therapy. After the therapy, the patients The datasets for this manuscript are not publicly available also presented a significant reduction in tremor (FTRS), spasticity because: Patients’ data need to be treated according to current (MAS), and increase on the grasp ability (BBT) and in the data protection laws and ethical guidelines. Requests to access the cognitive state (MOCA). datasets should be directed to MS-R, sebastian@gtec.at. In general, the first sign that patients reported during the therapy was a reduction in spasticity, followed with improvement in motor function. The reduced spasticity may drive the ETHICS STATEMENT improved range of motion and reduced tremor, something that can explain the improvements on FTRS, BBT and FMAue. The studies involving human participants were reviewed and The MOCA scale also showed significant improvement, which approved by Ethikkommission des Landes Oberösterreich. The may be related to the need for concentration during the BCI patients/participants provided their written informed consent to sessions in order to get positive feedback. The patients have to participate in this study. learn to maintain concentration during the sessions to improve their motor skills using BCI. AUTHOR CONTRIBUTIONS GENERAL DISCUSSION MS-R participated in the data acquisition, performed the analysis, and did the main contribution to the manuscript writing. EU The BSI parameter can be calculated in real-time using portable supervised the data analysis and assisted in the manuscript and practical EEG tools, and thus could be used during stroke preparation. RO supervised the whole process, data acquisition, diagnosis or ongoing monitoring of patients’ brain activity during analysis, and manuscript revision. JD-F supervised the signal stroke rehabilitation and recovery. More broadly, the BSI and LC processing methods. NM participated to the data acquisition parameters might contribute to other neurological assessments from stroke participants. CM-P did the acquisition of healthy and ongoing monitoring of brain damage and recovery. subjects. SS participated in the data analysis. BA provided One limitation of this study is the absence of a healthy group scientific input and contributed to the manuscript writing. CG that performed the same BCI training as the patient group, which supervised the whole project and reviewed the manuscript. WC prevents us from comparing LC between these groups. The study contributed to the data analysis. may also be limited by the unequal numbers of participants across the three stroke subgroups, and additional work is needed to identify any age differences between the control and stroke FUNDING groups. Overcoming these limitations will require substantial additional work with more participants in a larger study, which We appreciate the collaboration of the Government of Catalonia we are currently exploring. Additionally, a great number of that partially supported this study with the Industrial Doctorates hypotheses tests were carried out in the present study. Therefore, Plan (2017/DI/065). This study was also supported by the EEG- results referred to as “statistically significant” may have been DDS project (E! 12274 – EUROSTARS), the SeizureAI project obtained by chance and we strongly recommend considering (E! 12946 – EUROSTARS), and MSCA-RISE grant Pro-Gait the obtained p-values and effect sizes when interpreting the (No. 778043). results of this study. In addition to broader work with more participants, future research could: explore variants of the different measures that SUPPLEMENTARY MATERIAL we used that might be more informative; identify correlations with other types of diagnoses and therapies relating to motor The Supplementary Material for this article can be found (and perhaps other) impairment and recovery; evaluate these and online at: https://www.frontiersin.org/articles/10.3389/fnins. other parameters in tandem with other methods to treat stroke, 2020.00582/full#supplementary-material REFERENCES Bender, R., and Lange, S. (2001). Adjusting for multiple testing—when and how? J. Clin. Epidemiol. 54, 343–349. doi: 10.1016/S0895-4356(00)00314-0 Agius Anastasi, A., Falzon, O., Camilleri, K., Vella, M., and Muscat, R. (2017). Brain Biasiucci, A., Leeb, R., Iturrate, I., Perdikis, S., Al-Khodairy, A., Corbet, T., et al. symmetry index in healthy and stroke patients for assessment and prognosis. (2018). Brain-actuated functional electrical stimulation elicits lasting arm motor Stroke Res. Treat. 2017, 1–9. doi: 10.1155/2017/8276136 recovery after stroke. Nat. Commun. 9:2421. doi: 10.1038/s41467-018-04673-z Frontiers in Neuroscience | www.frontiersin.org 14 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation Carlson, T., and del, R. M. J. (2013). Brain-controlled wheelchairs: a robotic Kim, J., Thayabaranathan, T., Donnan, G. A., Howard, G., Howard, V. J., architecture. IEEE Robot. Autom. Mag. 20, 65–73. doi: 10.1109/MRA.2012. Rothwell, P. M., et al. (2020). Global Stroke Statistics 2019. Int. J. Stroke 2229936 139:174749302090954. doi: 10.1177/1747493020909545 Cervera, M. A., Soekadar, S. R., Ushiba, J., del, R. M. J., Liu, M., Birbaumer, N., Kohr, R. L., and Games, P. A. (1974). Robustness of the analysis of variance, the et al. (2018). Brain-computer interfaces for post-stroke motor rehabilitation: welch procedure and a box procedure to heterogeneous variances. J. Exp. Educ. a meta-analysis. Ann. Clin. Transl. Neurol. 5, 651–663. doi: 10.1002/ 43, 61–69. doi: 10.1080/00220973.1974.10806305 acn3.544 Krauss, P., Metzner, C., Schilling, A., Tziridis, K., Traxdorf, M., Wollbrink, A., Cho, W., Guger, C., Heilinger, A., Ortner, R., Murovec, N., Xu, R., et al. (2018). et al. (2018). A statistical method for analyzing and comparing spatiotemporal “Motor rehabilitation for hemiparetic stroke patients using a brain-computer cortical activation patterns. Sci. Rep. 8:5433. doi: 10.1038/s41598-018-23765-w interface method,” in Proceedings of the 2018 IEEE International Conference Leon-Carrion, J., Martin-Rodriguez, J. F., Damas-Lopez, J., Barroso y Martin, J. M., on Systems, Man, and Cybernetics (SMC), (Piscataway, NJ: IEEE), 1001–1005. and Dominguez-Morales, M. R. (2009). Delta–alpha ratio correlates with level doi: 10.1109/SMC.2018.00178 of recovery after neurorehabilitation in patients with acquired brain injury. Cho, W., Sabathiel, N., Ortner, R., Lechner, A., Irimia, D. C., Allison, B. Z., Clin. Neurophysiol. 120, 1039–1045. doi: 10.1016/j.clinph.2009.01.021 et al. (2016). Paired associative stimulation using brain-computer interfaces for Mäkelä, J. P., Lioumis, P., Laaksonen, K., Forss, N., Tatlisumak, T., Kaste, M., stroke rehabilitation: a pilot study. Eur. J. Transl. Myol. 26:6132. doi: 10.4081/ et al. (2015). Cortical excitability measured with nTMS and MEG during Stroke ejtm.2016.6132 recovery. Neural Plast. 2015, 1–8. doi: 10.1155/2015/309546 Dobkin, B. H. (2007). Brain-computer interface technology as a tool to augment McFarland, D. J., and Wolpaw, J. R. (2017). EEG-based brain–computer interfaces. plasticity and outcomes for neurological rehabilitation. J. Physiol. 579, 637–642. Curr. Opin. Biomed. Eng. 4, 194–200. doi: 10.1016/j.cobme.2017.11.004 doi: 10.1113/jphysiol.2006.123067 Muniz, C. F., Shenoy, A. V., O’Connor, K. L., Bechek, S. C., Boyle, E. J., Faust, O., Acharya, U. R., Adeli, H., and Adeli, A. (2015). Wavelet-based EEG Guanci, M. M., et al. (2016). Clinical development and implementation of processing for computer-aided seizure detection and epilepsy diagnosis. Seizure an institutional guideline for prospective EEG monitoring and reporting of 26, 56–64. doi: 10.1016/j.seizure.2015.01.012 delayed cerebral ischemia. J. Clin. Neurophysiol. 33, 217–226. doi: 10.1097/ Foreman, B., and Claassen, J. (2012). Quantitative EEG for the detection of brain WNP.0000000000000281 ischemia. Crit. Care 16:216. doi: 10.1186/cc11230 Neuper, C., Wörtz, M., and Pfurtscheller, G. (2006). “ERD/ERS patterns reflecting Gladstone, D. J., Danells, C. J., and Black, S. E. (2002). The Fugl-meyer assessment sensorimotor activation and deactivation,” in Progress in Brain Research, eds S. of motor recovery after stroke: a critical review of its measurement properties. Waxman, D. G. Stein, D. Swaab, and H. Fields, (Amsterdam: Elsevier), 211–222. Neurorehabil. Neural Repair 16, 232–240. doi: 10.1177/15459680240110 doi: 10.1016/S0079-6123(06)59014-4 5171 Nishida, K., Yoshimura, M., Isotani, T., Yoshida, T., Kitaura, Y., Saito, A., et al. Graimann, B., Huggins, J., Levine, S., and Pfurtscheller, G. (2002). Visualization (2011). Differences in quantitative EEG between frontotemporal dementia and of significant ERD/ERS patterns in multichannel EEG and ECoG data. Clin. Alzheimer’s disease as revealed by LORETA. Clin. Neurophysiol. 122, 1718– Neurophysiol. 113, 43–47. doi: 10.1016/S1388-2457(01)00697-6 1725. doi: 10.1016/j.clinph.2011.02.011 Hurlbert, S. H., and Lombardi, C. M. (2012). Lopsided reasoning on lopsided tests Pfurtscheller, G., and Aranibar, A. (1979). Evaluation of event-related and multiple comparisons. Aust. New Zeal. J. Stat. 54, 23–42. doi: 10.1111/j. desynchronization (ERD) preceding and following voluntary self- 1467-842X.2012.00652.x paced movement. Electroencephalogr. Clin. Neurophysiol. 46, 138–146. Ikkai, A., Dandekar, S., and Curtis, C. E. (2016). Lateralization in Alpha-band doi: 10.1016/0013-4694(79)90063-4 oscillations predicts the locus and spatial distribution of attention. PLoS One Piano, C., Mazzucchi, E., Bentivoglio, A. R., Losurdo, A., Calandra Buonaura, G., 11:e0154796. doi: 10.1371/journal.pone.0154796 Imperatori, C., et al. (2017). Wake and sleep EEG in patients with Huntington Irimia, D., Sabathiel, N., Ortner, R., Poboroniuc, M., Coon, W., Allison, B. Z., disease. Clin. EEG Neurosci. 48, 60–71. doi: 10.1177/1550059416632413 et al. (2016). “recoveriX: a new BCI-based technology for persons with stroke,” Pichiorri, F., Morone, G., Petti, M., Toppi, J., Pisotta, I., Molinari, M., et al. in Proceedings of the 2016 38th Annual International Conference of the IEEE (2015). Brain-computer interface boosts motor imagery practice during stroke Engineering in Medicine and Biology Society (EMBC, (Piscataway, NJ: IEEE), recovery. Ann. Neurol. 77, 851–865. doi: 10.1002/ana.24390 1504–1507. doi: 10.1109/EMBC.2016.7590995 Prabhakaran, S., Zarahn, E., Riley, C., Speizer, A., Chong, J. Y., Lazar, R. M., Irimia, D. C., Cho, W., Ortner, R., Allison, B. Z., Ignat, B. E., Edlinger, G., et al. (2008). Inter-individual variability in the capacity for motor recovery et al. (2017). Brain-computer interfaces with multi-sensory feedback for stroke after ischemic stroke. Neurorehabil. Neural Repair 22, 64–71. doi: 10.1177/ rehabilitation: a case study. Artif. Organs 41, E178–E184. doi: 10.1111/aor. 1545968307305302 13054 Rabiller, G., He, J. W., Nishijima, Y., Wong, A., and Liu, J. (2015). Perturbation Irimia, D. C., Ortner, R., Poboroniuc, M. S., Ignat, B. E., and Guger, C. (2018). High of brain oscillations after Ischemic stroke: a potential biomarker for post- classification accuracy of a motor imagery based brain-computer interface for stroke function and therapy. Int. J. Mol. Sci. 16, 25605–25640. doi: 10.3390/ stroke rehabilitation training. Front. Robot. AI 5. doi: 10.3389/frobt.2018.00130 ijms161025605 Janssen, T. W. P., Hillebrand, A., Gouw, A., Geladé, K., Van Mourik, R., Maras, A., Ramos-Murguialday, A., Broetz, D., Rea, M., Läer, L., Yilmaz, Ö, Brasil, F. L., et al. et al. (2017). Neural network topology in ADHD; evidence for maturational (2013). Brain-machine interface in chronic stroke rehabilitation: a controlled delay and default-mode network alterations. Clin. Neurophysiol. 128, 2258– study. Ann. Neurol. 74, 100–108. doi: 10.1002/ana.23879 2267. doi: 10.1016/j.clinph.2017.09.004 Remsik, A., Young, B., Vermilyea, R., Kiekhoefer, L., Abrams, J., Evander Elmore, Jeong, D. H., Kim, Y., Do Song, I. U., Chung, Y. A., and Jeong, J. (2016). Wavelet S., et al. (2016). A review of the progression and future implications of brain- energy and wavelet coherence as eeg biomarkers for the diagnosis of Parkinson’s computer interface therapies for restoration of distal upper extremity motor disease-related dementia and Alzheimer’s disease. Entropy 18:8. doi: 10.3390/ function after stroke. Expert Rev. Med. Devices 13, 445–454. doi: 10.1080/ e18010008 17434440.2016.1174572 Kaiser, V., Daly, I., Pichiorri, F., Mattia, D., Müller-Putz, G. R., and Neuper, C. Shokur, S., Donati, A. R. C., Campos, D. S. F., Gitti, C., Bao, G., Fischer, D., (2012). Relationship between electrical brain responses to motor imagery and et al. (2018). Training with brain-machine interfaces, visuo-tactile feedback and motor impairment in stroke. Stroke 43, 2735–2740. doi: 10.1161/STROKEAHA. assisted locomotion improves sensorimotor, visceral, and psychological signs in 112.665489 chronic paraplegic patients. PLoS One 13:e0206464. doi: 10.1371/journal.pone. Kanda, P. A., de, M., Anghinah, R., Smidth, M. T., and Silva, J. M. (2009). The 0206464 clinical use of quantitative EEG in cognitive disorders. Dement. Neuropsychol. Tomarken, A. J., and Serlin, R. C. (1986). Comparison of ANOVA alternatives 3, 195–203. doi: 10.1590/S1980-57642009DN30300004 under variance heterogeneity and specific noncentrality structures. Psychol. Kiiski, H., Rueda-Delgado, L. M., Bennett, M., Knight, R., Rai, L., Roddy, D., Bull. 99, 90–99. doi: 10.1037/0033-2909.99.1.90 et al. (2020). Functional EEG connectivity is a neuromarker for adult attention van Putten, M. J. A. M. (2006). Extended BSI for continuous EEG monitoring deficit hyperactivity disorder symptoms. Clin. Neurophysiol. 131, 330–342. doi: in carotid endarterectomy. Clin. Neurophysiol. 117, 2661–2666. doi: 10.1016/ 10.1016/j.clinph.2019.08.010 j.clinph.2006.08.007 Frontiers in Neuroscience | www.frontiersin.org 15 July 2020 | Volume 14 | Article 582 Sebastián-Romagosa et al. qEEG Biomarkers for Stroke Rehabilitation van Putten, M. J. A. M. (2007). The revised brain symmetry index. Clin. Conflict of Interest: MS-R, RO, JD-F, CM-P, and CG are employed at g.tec Neurophysiol. 118, 2362–2367. doi: 10.1016/j.clinph.2007.07.019 medical engineering Spain SL and CG is CEO of g.tec medical engineering Spain van Putten, M. J. A. M., Peters, J. M., Mulder, S. M., de Haas, J. A. M., Bruijninckx, SL and g.tec medical engineering GmbH. This study was conducted primarily C., and Tavy, D. L. J. (2004). A brain symmetry index (BSI) for online EEG as a component of a Ph.D. program, and our motivation for conducting the monitoring in carotid endarterectomy. Clin. Neurophysiol. 115, 1189–1194. study to explore different methods was entirely scientific. However, the described doi: 10.1016/j.clinph.2003.12.002 algorithms may be commercialized by g.tec medical engineering GmbH in the Wang, J., Barstein, J., Ethridge, L. E., Mosconi, M. W., Takarae, Y., and Sweeney, future. J. A. (2013). Resting state EEG abnormalities in autism spectrum disorders. J. Neurodev. Disord. 5:24. doi: 10.1186/1866-1955-5-24 The remaining authors declare that the research was conducted in the absence of Wolpaw, J. R. (2007). Brain-computer interfaces as new brain output pathways. any commercial or financial relationships that could be construed as a potential J. Physiol. 579, 613–619. doi: 10.1113/jphysiol.2006.125948 conflict of interest. Woytowicz, E. J., Rietschel, J. C., Goodman, R. N., Conroy, S. S., Sorkin, J. D., Whitall, J., et al. (2017). Determining levels of upper extremity movement Copyright © 2020 Sebastián-Romagosa, Udina, Ortner, Dinarès-Ferran, Cho, impairment by applying a cluster analysis to the Fugl-Meyer assessment of Murovec, Matencio-Peralba, Sieghartsleitner, Allison and Guger. This is an open- the upper extremity in chronic stroke. Arch. Phys. Med. Rehabil. 98, 456–462. access article distributed under the terms of the Creative Commons Attribution doi: 10.1016/j.apmr.2016.06.023 License (CC BY). The use, distribution or reproduction in other forums is permitted, Wu, J., Srinivasan, R., Burke Quinlan, E., Solodkin, A., Small, S. L., and provided the original author(s) and the copyright owner(s) are credited and that the Cramer, S. C. (2016). Utility of EEG measures of brain function in patients original publication in this journal is cited, in accordance with accepted academic with acute stroke. J. Neurophysiol. 115, 2399–2405. doi: 10.1152/jn.00978. practice. No use, distribution or reproduction is permitted which does not comply 2015 with these terms. Frontiers in Neuroscience | www.frontiersin.org 16 July 2020 | Volume 14 | Article 582