ORIGINAL RESEARCH published: 25 February 2022 doi: 10.3389/fneur.2022.838170 Resting-State Beta-Band Recovery Network Related to Cognitive Improvement After Stroke Sandra Pusil 1*† , Lucía Torres-Simon 1† , Brenda Chino 2 , María Eugenia López 1 , Leonides Canuet 1 , Álvaro Bilbao 3 , Fernando Maestú 1 and Nuria Paúl 1 1 Department of Experimental Psychology, Universidad Complutense de Madrid, Madrid, Spain, 2 Institute of Neuroscience, Autonomous University of Barcelona, Barcelona, Spain, 3 National Centre for Brain Injury Treatment, Centro de Referencia Estatal de Atención Al Daño Cerebral (CEADAC), Madrid, Spain Edited by: Background: Stroke is the second leading cause of death worldwide and it causes Rafeed Alkawadri, important long-term cognitive and physical deficits that hamper patients’ daily activity. University of Pittsburgh Medical Neuropsychological rehabilitation (NR) has increasingly become more important to Center, United States recover from cognitive disability and to improve the functionality and quality of life Reviewed by: Shennan Aibel Weiss, of these patients. Since in most stroke cases, restoration of functional connectivity SUNY Downstate Medical Center, (FC) precedes or accompanies cognitive and behavioral recovery, understanding the United States Donna Clark Tippett, electrophysiological signatures underlying stroke recovery mechanisms is a crucial Johns Hopkins Medicine, scientific and clinical goal. United States Methods: For this purpose, a longitudinal study was carried out with a sample *Correspondence: Sandra Pusil of 10 stroke patients, who underwent two neuropsychological assessments and two spusil@ucm.es resting-state magnetoencephalographic (MEG) recordings, before and after undergoing † These authors have contributed a NR program. Moreover, to understand the degree of cognitive and neurophysiological equally to this work and share first impairment after stroke and the mechanisms of recovery after cognitive rehabilitation, authorship stroke patients were compared to 10 healthy controls matched for age, sex, and Specialty section: educational level. This article was submitted to Findings: After intra and inter group comparisons, we found the following results: Applied Neuroimaging, a section of the journal (1) Within the stroke group who received cognitive rehabilitation, almost all cognitive Frontiers in Neurology domains improved relatively or totally; (2) They exhibit a pattern of widespread increased Received: 17 December 2021 in FC within the beta band that was related to the recovery process (there were no Accepted: 03 February 2022 Published: 25 February 2022 significant differences between patients who underwent rehabilitation and controls); (3) Citation: These FC recovery changes were related with the enhanced of cognitive performance. Pusil S, Torres-Simon L, Chino B, Furthermore, we explored the capacity of the neuropsychological scores before López ME, Canuet L, Bilbao Á, rehabilitation, to predict the FC changes in the brain network. Significant correlations Maestú F and Paúl N (2022) Resting-State Beta-Band Recovery were found in global indexes from the WAIS-III: Performance IQ (PIQ) and Perceptual Network Related to Cognitive Organization index (POI) (i.e., Picture Completion, Matrix Reasoning, and Block Design). Improvement After Stroke. Front. Neurol. 13:838170. Keywords: stroke, functional connectivity (FC), MEG (magnetoencephalography), cognitive performance, doi: 10.3389/fneur.2022.838170 neuropsychological rehabilitation Frontiers in Neurology | www.frontiersin.org 1 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke INTRODUCTION or Alzheimer’s disease (24). Moreover, MEG has previously been used to demonstrate the disruption and recovery of functional Stroke is considered the second leading cause of death and networks, and even its relationship with cognitive improvement the third leading cause of disability worldwide (1). It is a after undergoing a NR program in acquired brain pathologies heterogeneous pathology with diverse clinical manifestations due such as stroke (22, 25, 26) or traumatic brain injury (TBI) to its possible etiologies (i.e., hemorrhagic, or ischemic), locations (13, 27). Thanks to the relevance of the neurophysiological (i.e., different vascular vessels or arteries), and size of the lesion (2, changes found in previous literature, it seems plausible that MEG 3). However, most stroke survivors suffer from different degrees may provide interesting information about how NR may induce of cognitive disabilities (4–6). These patients may have damage specific cognitive recovery in stroke patients. in general cognitive performance with important functional According to the aforementioned antecedents, the aims of the disability, which has been broadly reported in the scientific present exploratory study are: (1) to understand the cognitive literature (7–9). Although stroke tends to impact on attention improvement achieved in stroke patients that received NR; (2) to and executive function compared with its impact on memory, explore the possible neurophysiological mechanisms underlying a malfunction in these cognitive domains could worsen the the recovery process, by using FC on frequency bands obtained performance in other cognitive areas (3, 4, 6, 10). In any case, with MEG; and (3) to evaluate if these neurophysiological it is important to highlight the role of non-pharmacological changes are related with cognitive improvement. For this rehabilitation especially neuropsychological rehabilitation (NR) purpose, we carried out a longitudinal study in which a sample in order to improve cognitive abilities and daily functions (10, of 10 stroke patients (stroke patients) were examined at two 11). Neuropsychological rehabilitation is a systematic therapeutic different time points. The first was before NR (from now on activity oriented functionally based on the assessment and we will say pre-condition), and the second was after NR (from understanding of the cognitive deficits, emotional disturbances, now on we will say post-condition). At both time points patients disruptive behaviors, and functional disorders of patients (12, were cognitively evaluated and underwent resting-state MEG 13), and includes interventions that might be compensatory, recordings. Moreover, to understand the degree of cognitive educational, or restorative (10). and neurophysiological disruption after stroke, and the recovery According to some authors and approaches post-stroke mechanisms in stroke patients who were enrolled on the NR, data deficits have long been considered to be fundamentally associated for a control group were included with 10 healthy controls paired with the location of the lesion (14). This could be particularly true in age, sex, and educational level. for sensorimotor or language deficits, which are closely related to the damage to the specific eloquent cortex. However, it has been shown that although structural damage from stroke is usually MATERIALS AND METHODS focal, remote disturbances may occur in brain distant regions from the primary area of damage (15, 16). This phenomenon Participants was previously associated with the concept of diaschisis, but The total dataset consisted of 20 subjects: 10 stroke patients (2 it is currently explained by the disruption of structural and females/8 males; mean age 44.9 ± 8.94; mean level of education functional connectivity (FC) between brain areas (17). This way 4.44 ± 0.97) and 10 healthy controls (2 females/8 males; mean of understanding the functioning of the brain gives a crucial age 43 ± 12.72; mean level of education 4.78 ± 0.67). The role to NR in the process of cognitive recovery, since it allows a mean time from the onset of the stroke to the start of the holistic management of cognitive impairment in contrast to other study was 6.3 months, and the rehabilitation program lasted 7 more goal-oriented therapies. months. The patient’s lesions were both ischemic (i.e., infarction; In this context, as the restoration of FC precedes or n = 5) and hemorrhagic (i.e., intracerebral hemorrhage; n = 5) accompanies in most cases, cognitive and behavioral recovery in and the stroke was located in different brain areas (for patient stroke patients (18, 19), understanding the electrophysiological detailed descriptive data see Table 1). To be enrolled in the study, signatures underlying stroke recovery mechanisms is a crucial patients had to be diagnosed with a first-ever stroke, showing scientific goal. This information could help the clinical a compatible lesion observed on computerized tomography community to anticipate and modify NR programs to achieve (CT) or magnetic resonance imaging (MRI). Although initially, a more effective cognitive recovery, and consequently, improve after the stroke some patients showed loss of consciousness [as patients’ quality of life. With this purpose, in the present study we reported in Table 1 with the Glasgow Coma Scale (28)], at the used the magnetoencephalography (MEG), a neurophysiological beginning of the study all patients were neurologically stable technique that allows a comprehensive analysis of brain dynamics without alterations in consciousness or alertness, and none of (20, 21). While the functional magnetic resonance image (fMRI) them showed epileptiform discharges on MEG recordings. is intrinsically limited by the hemodynamic response, MEG Exclusion criteria were the following: a stroke involving directly measures cortical neural activity. That means that the the brainstem or cerebellum, a diagnosis of neurological or modified vasomotor reactivity and neurovascular uncoupling in psychiatric diseases other than stroke, and a history of TBI, drug, stroke easily affects the blood oxygen level-dependent (BOLD) or alcohol abuse. response but leaves the MEG signal intact (22). The study of MEG Patients were recruited from the National Brain Injury signatures is well-established for early detection and prognosis Rehabilitation Center and from Lescer Brain Injury in neurodegenerative disorders, such as multiple sclerosis (23) Rehabilitation Center (Madrid, Spain), and all of them were Frontiers in Neurology | www.frontiersin.org 2 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke TABLE 1 | Clinical and sociodemographic characteristics of the patients. Patient Age Sex Education GCS Stroke etiology Stroke lesion 1 44 M 3 12 Ischemia Right fronto-parietal 2 45 F 5 7 Ischemia Right middle cerebral artery 3 47 M 5 9 Ischemia Left middle cerebral artery 4 47 M 4 12 Ischemia Right middle and anterior cerebral arteries 5 60 M 3 12 Ischemia Left middle cerebral artery 6 28 F 4 7 Hemorrhage Thalamus and left basal ganglia 7 35 M 4 8 Hemorrhage Right intraparenchymal 8 41 M 5 7 Hemorrhage Left basal ganglia 9 49 M 6 9 Hemorrhage Right basal ganglia 10 53 M 5 9 Hemorrhage Left thalamus N = 10 44.9 ± 8.9 8 M/2 F 4.4 ± 0.9 9.2 ± 2.1 5 isch/5 hem Education (1, illiterate/functional illiterate; 2, elemental studies; 3, school graduate; 4, high school studies; 5, university graduate studies; 6, university post-graduate studies). enrolled in a NR program. Healthy controls were matched with Neuropsychological Rehabilitation patients for age, sex, and education level, and they did not have a Program previous history of psychiatric or neurological disorders. All stroke patients received an integrated treatment based on As previously mentioned, patients underwent MEG the holistic-comprehensive model proposed by Ben-Yishay and recordings and neuropsychological evaluation in two different Diller (41). This program consists of 1 h/day of occupational moments: (1) Pre-condition (Pre): at the beginning of the therapy, 1 h/2 days/week of neuropsychological therapy, and study, before NR program; and (2) Post-condition (Post): after 2 h/day of group cognitive therapy (memory and executive completing the NR program. In the case of healthy controls, both function/social skills). Neuropsychological therapy aimed to data, neuropsychological and neurophysiological, were obtained improve attention, working memory, learning, memory and only once, at the beginning of the study. problem solving/executive functions, and emotional-behavioral problems, through evidence-based techniques that included both restorative and compensatory strategies. Neuropsychological Ethics Statement treatment goals in each case were defined to achieve maximum Methods were carried out in accordance with approved cognitive independence in daily living. In addition, patients guidelines and regulations. The study was approved by the underwent 1 h of physiotherapy and half an hour of speech National Brain Injury Rehabilitation Center Ethics Committee therapy, in those cases that needed it. This rehabilitation plan (Madrid), and all participants or legal representatives signed a met the following requirements: (1) agreed by the family written informed consent prior to participation. and all professionals involved; (2) formulated in a specific and operational manner (3) focused on meaningful goals for the patients that allow them to achieve greater personal Neuropsychological Assessment autonomy, community integration, and adaptation to their All participants underwent a comprehensive neuropsychological deficits; and (4) reviewed monthly. In addition, all patients evaluation with the aim to identify their cognitive status attended psychotherapy sessions to help them in the process of in multiple cognitive domains (attention, memory, language, accepting their new situation. executive functions, and visuospatial abilities) as well as their functional performance. The extensive neuropsychological Magnetoencephalographic Recordings assessment included: the Wechsler Adult Intelligence Scale III Magnetic fields were recorded using a 148-channel whole-head (WAIS III) (29), the Brief Test of Attention (BTA) (30), the magnetometer (4D-MAGNES_2500 WH, 4-D Neuroimaging) Trail Making Test (TMT) (31), the Stroop Color Word Test confined in a magnetically shielded room at the Universidad (32, 33), the Wisconsin Card Sorting Test (WCST) (34), the Complutense of Madrid (Spain). Fields were measured during a Tower of Hanoi (35), the Zoo Map Test [from the Behavioral 2-min resting-state eyes-closed condition and were sampled at a Assessment of the Dysexecutive Syndrome (36)], the Boston frequency rate of 618.17 Hz. Ocular, cardiac, muscular, and jump Naming Test (BNT) (37), the Digit Span Test [Wechsler Memory artifacts were identified first, by a visual inspection of an expert in Scale III (29)], the Visual Span Test [WMS-III; (29)], Logical MEG, and then removed using ICA (42) in Brainstorm software Memory and Visual Reproduction [WMS-III (29)], the Phonemic (43). Then, clean data were segmented into 4 s trial length, with and Semantic Fluency [Controlled Oral Word Association Test, a minimum of 20 artifact-free segments for each subject. The COWAT (38)], the Five Digit Test [FDT (39)], the Dysexecutive MEG data were filtered in the classical frequency bands: delta Questionnaire [DEX (36)], and the Patient Competency Rating (2–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), beta (12–30 Hz), and Scale [PCRS (40)]. gamma (30–45 Hz) for further analysis. Frontiers in Neurology | www.frontiersin.org 3 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke Source Reconstruction and Connectivity and Wilcoxon paired test for within-group comparison (Pre vs. Analysis Post conditions in the stroke patients’ group). In the case of To reconstruct MEG sources, we used the default anatomy neuropsychological variables, significant results were considered (15,000 vertices) of the MNI/Colin27 brain (44) in Brainstorm. with a p-value < 0.05 after applying false discovery rate (FDR) This template was warped according to the polhemus points corrected for multiple comparisons. For FC data, a total of (nasion and both preauricular) acquired during the head 10,000 permutations were used for each significant FC link, digitalization to obtain a better approximation of the real and results were considered significant with a p-value <0.005 shape of the subject’s head. The overlapping sphere model after applying FDR (51). Finally, with the aim to explore the was calculated as the forward modeling of MEG measures. relationships between FC and cognition, Spearman’s correlation Next, a noise covariance matrix was calculated to estimate analysis was employed. For all analyses the Matlab Statistical noise level in the MEG recordings. Sources were reconstructed Toolbox was used. using the weighted Minimum Norm Estimation (wMNE) (45). Weighted Minimum Norm Estimation is well-suited for the estimation of large-scale FC networks, since it addresses the RESULTS problem of volume conduction, reducing the correlations of spurious signals (46, 47). Magnetoencephalography sources Cognitive Changes After were grouped into 68 anatomical regions of interest (ROI) Neuropsychological Rehabilitation based on Brainstorm atlas Desikan-Killiany (48). For more As described before, the patients underwent a comprehensive details about the brain areas used, referred to Supplementary neuropsychological evaluation before and after the NR program. Material for Supplementary Table 1. We selected the mean as the From the total of battery tests, those scores with at least nine representative time series for each brain area delimited with the reported patients (46 scores in total) were included for the aforementioned atlas. statistical analysis. Pre-condition results indicated that stroke Functional connectivity was assessed using the corrected patients performed significantly worse compared with healthy version of the imaginary phase locking value (ciPLV), a phase controls in all cognitive domains (p < 0.05). Comparing pre synchronization measure that evaluates the distribution of phase and post conditions in the stroke patients’ group, results showed differences extracted from each of two sensor time series (49, 50). an important cognitive improvement with 33 scores (72% of Corrected version of the imaginary phase locking value (Equation the 46 total scores) significantly different between conditions. 1) was proposed by Bruña et al. (50) to remove the contribution Of these, 21 scores (46% of the 46 total scores), could be of the zero phase differences of PLV. Thus, this measure is considered relatively enhanced since in the post-condition, they insensitive to zero-lag effects, and it is corrected to remove the were significantly different to those corresponding to the healthy instantaneous phase contribution, which could be mainly due to controls (see Figure 2). The remaining 12 scores (26% of the 46 volume conduction. total scores) from the post-condition did not show significant differences with the healthy control group, indicating a total 1 −i(φX (t)−φY (t)) } improvement. To simplify the interpretation of these results, all T I{e ciPLVX,Y (t) = q (1) scores were clustered into several aggregated groups depending 2 1 − ( T1 R{e−i(φX (t)−φY (t)) }) on different cognitive domains: Functional performance, 4 scores (of DEX and PCRS); Executive Functions, 10 scores (of WCST, where ϕx and ϕy represent the phases of each of the two-time Tower of Hanoi, FDT and TMT); Attention, 1 score (of Brief series and stands for the imaginary part of the numerator and Test of Attention); Language, 2 scores (of BNT and FAS); the real part in the denominator. See Figure 1 for the analysis Episodic Memory, 4 scores (of WMS-III) and Working Memory, flow chart of the MEG data. 9 scores (of WMS-III and WAIS-III). The last three columns of Figure 2 correspond to the cognitive index of WAIS-III, Statistical Analyses including all their subtests: Verbal Comprehension Index (VCI, This study aims to find the possible neurophysiological substrates 4 scores); Processing Speed Index (PSI, 3 scores), and Perceptual of the recovery network underlying the cognitive enhancement Organization Index (POI, 4 scores). In addition, the three found stroke patient’s sample in the post condition (after NR), general indices of WAIS-III were also included [Verbal IQ, by using cognitive tests, functional scales, and FC measures. Performance IQ (PIQ), and Full-Scale IQ], as well as subtests In this context, we performed exploratory analyses with the Picture Arrangement and Comprehension. In summary, we data obtained from stroke patients and healthy controls. The found that all cognitive domains of the stroke patients’ group analysis of demographic data showed that there were no were fully or partially enhanced in the post-condition (see statistical differences in age, sex, and level of education between Figure 2). It is important to note that the VCI (that includes patients and controls (p > 0.05), so we did not include the WAIS-III index and all its subtests), was a cognitive domain them as confounding variables for the following explorations. without improvement in any of the four measures; and the four Non-parametric tests were used for all comparisons because scores maintained significant differences when comparing the variables were non-normally distributed and because of the small second scores of stroke patients with the healthy controls scores. sample size. Specifically, the Mann-Whitney U test was used for However, it is important to know that this index, already in the between groups analyses (stroke patients vs. healthy controls) pre-condition, shows a normal score (103.5) unlike the results of Frontiers in Neurology | www.frontiersin.org 4 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke FIGURE 1 | (Top) Sample description (stroke patients and healthy control groups). Description of the protocol followed by each group and the variables collected for this study. (Bottom) MEG Analysis flowchart. Sequential pipeline of the analysis performed on the MEG data. the other cognitive index of the WAIS-III (WMI 91.9; PSI 80.3; FC within the beta band was found. Stroke patients showed PRI 82.4) and the others all cognitive tests. In addition, their significantly (p < 0.005) higher FC in the post-condition result in the post-condition was 105, although it continues to be compared to the pre-condition in a variety of links comprising statistically different from the healthy control group (117.6). intra and inter-hemispheric, and antero-posterior long-range connections (Figure 4). Moreover, when assessing the possible differences in FC Functional Connectivity Disruption After between groups (post-condition and healthy controls) we did not Stroke: Differences Between Stroke find any statistically significant differences. This result indicated Patients and Healthy Controls that the original FC disruption in the beta band was restored in In order to assess the possible disruption of the patients’ stroke patients who went through the NR. network due to the stroke, their FC in the pre-condition was No significant results were found in other frequency bands in compared with the FC of the healthy controls. Stroke patients the pre and post comparison after FDR correction (p < 0.005). exhibited significant FC reduction in the beta band (p < 0.005, Nevertheless, there is a clear pattern of enhanced connectivity FDR corrected) that comprised intra and inter-hemispheric in low frequency bands (delta and theta) in the pre stage when connections (Figure 3). No significant results were found in compared with the brain activity of stroke patients recorded other frequency bands. after the rehabilitation when a less restrictive statistical threshold was used (p < 0.05). Detailed description of these results could be found in the Supplementary Material, Appendix 3. Functional Connectivity After These results were not included in the main findings of the Rehabilitation: The Recovery Network present study because we wanted to focus on the most reliable When assessing the possible FC differences between stroke FC signature, keeping the p < 0.005 value as the go/no go patients’ conditions, a clear pattern of widespread increased statistical limit. Frontiers in Neurology | www.frontiersin.org 5 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke FIGURE 2 | Neuropsychological tests’ changes after NR. Functional Performance (4 scores: DEX and PCRS); Executive Functions (10 scores: WCST, Tower of Hanoi, FDT, and TMT); Attention (1 score: Brief Test of Attention); Language (2 scores: BNT and FAS); Episodic Memory (4 scores: WMS-III); Working Memory (9 scores: WAIS-III and WMS-III). The last three columns correspond to cognitive indices of WAIS-III, including all their subtests: VCI (verbal comprehension, 4 scores); PSI (speed processing, 3 scores); and POI (perceptual organization, 4 scores). Percentages were calculated intra-domain. Test scores included. DEX, DEX-family, DEX-difference; PCRS, PCRS-family, PCRS-difference; WCST, WCST-categories, WCST-conceptual, WCST-persevering; Tower of Hanoi, TH-3D-time, TH-3D-movements, TH-4D-time, TH-4D-movements; Five Digit Test, FDT-switching, FDT-flexibility; Brief Test of Attention, BTA-total score; BNT, BNT-total score; FAS, FAS-total score; WMS-III-episodic-memory, WMS-III-logical memory 1, WMS-III-logical memory 2, WMS-III-visual reproduction 1, WMS-III-visual reproduction 2; WMS-working memory, WMS-III-forward digit span, WMS-III-backward digit span, WMS-III-forward visual span, WMS-III-backward visual span; WAIS-working memory, WAIS-digit span, WAIS-arithmetic, WAIS-letter number sequencing, WAIS-working memory index; WAIS-VCI, WAIS-vocabulary, WAIS-information, WAIS-similarities, WAIS-verbal comprehension index; WAIS-PSI, WAIS-symbol search, WAIS-digit symbol, WAIS-processing speed index; WAIS-POI, WAIS-block design, WAIS-matrix reasoning, WAIS-picture completion; WAIS-perceptual organization index. Correlations Between the Brain and the correlation analyses, the neuropsychological experts’ team Cognitive Recovery Patterns choose: Functional Performance (DEX-F), Executive functions With the aim to explore if FC changes were related with the (WCST-Persevering, Tower of Hanoi-3D-T), Attention (BTA), enhanced cognitive performance in the stroke patients’ group, Episodic memory (WMS-III-LM1), Working memory (Digit span we firstly calculated a ratio considering the strength of each test), and Language (BNT). Moreover, the WAIS-III general functional link that differed between both conditions (FC ratio indexes and some WAIS-III cognitive indexes were included = Post/Pre). Then, we averaged these FC link ratios in just (FIQ, PIQ, VIQ, PSI, POI). We finally included for the one value for each stroke patient. This provided a unique Spearman’s correlation analyses the average FC strength ratio FC marker for each patient that condensed the information and 12 neuropsychological scores differences, illustrative of the obtained by the whole network and the two MEG sessions. cognitive improvement, for each stroke patient. Next, for cognitive scores, we calculated the performance In order to facilitate the understanding of each patient differences (D) for the most representative tests of each cognitive improvement an extended material about individual neuropsychological domain between pre and post conditions neuropsychological performance was added in Appendix 2 in the stroke patients’ group (D = Post–Pre), with the aim in Supplementary Material. There, the punctuations of finding the strongest cognitive improvement, to reduce the for each patient before and after the rehabilitation are redundancy of the information (since several scores measured described in detail for those tests included in the present the similar aspects of the same cognitive domain) and to correlation analysis. avoid the statistical pitfall of multiple comparisons. Regarding We found three positive and significant recovery signatures the selection of the most representative scores included for correlations between the FC strength ratio and three cognitive Frontiers in Neurology | www.frontiersin.org 6 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke FIGURE 3 | (A) Significant FC results (p < 0.005, corrected) in the beta band (12–30 Hz) when comparing healthy controls vs. stroke patients in the pre-condition. Line thickness of significant links is proportional to FC values (a higher value corresponds to thicker lines, and vice versa); (B) Significant FC links in the beta band are represented as bar graphs. Red color represents higher connectivity values for healthy controls compared to stroke patients and blue color illustrates lower connectivity values for stroke patients compared to healthy controls. ROIs included: rmOrbG, Right Medial Orbito Frontal Gyrus; lmOrbG, Left Medial Orbito Frontal Gyrus; rMFG, Right Middle Frontal Gyrus; rOrbG, Lateral Orbito Frontal Gyrus; rIFGtri, Right Inferior Frontal Parstriangularis; lITG, Left Inferior Temporal Gyrus; rITG, Right Inferior Temporal Gyrus; lPHG, Left Parahippocampal Gyrus; rPHG, Right Parahippocampal Gyrus; rMCC, Right Posterior Cingulate Gyrus; rPoCG, Right Postcentral Gyrus; rSMG, Right Supramarginal Gyrus; rLING, Right Lingual Cortex; lMOG, Left Lateral Occipital Gyrus. measures: Full Scale IQ of WAIS-III (R = 0.833; p = 0.015), BNT Additionally, this study is focused on the exploration of the (R = 0.756; p = 0.035), and LM1 of WMS-III (R = 0.854; p = nature of the relationships between neurophysiological and 0.010) (Figure 5). neuropsychological changes. In this sense, our results indicate a positive effect in acute Prediction of Brain FC Recovery Based on stroke patients who received cognitive rehabilitation on both levels, the cognitive system and brain functioning. Nevertheless, Cognitive Performance After Stroke the lack of a clinical control group (i.e., stroke patients without Lastly, with the aim of exploring the predictive capacity of the rehabilitation) did not allow us to make causality assumptions, neuropsychological test scores and the brain network recovery, assuring that cognitive improvement is due specifically or we correlated the cognitive scores of the pre-condition and the uniquely to cognitive rehabilitation because it could also FC strength ratio (by using Spearman correlation analyses). represent some degree of spontaneous clinical recovery after Thus, we observed two markers for recovery prediction in stroke. Then, according to the results of the present study, stroke two global cognitive domains: (1) PIQ, with a significant positive patients who undertook rehabilitation significantly improved correlation between FC strength ratio and the PIQ scores (R = their performance in 72% of cognitive and functional scores. 0.850, p = 0.011); (2) Perceptual Organization, with a significant But we also found than in neuropsychological scores related to positive correlation between FC strength ratio and the POI scores specific cognitive domains such as executive functions, attention, (R = 0.874, p = 0.007). Furthermore, within POI, we found a language, episodic, and working memory, stroke patients showed positive association with Picture Completion (R = 0.732; p = a relative improvement (46% improved but there were significant 0.048), Matrix Reasoning (R = 0.795; p = 0.023), and Block differences between the post-condition and the control group), Design (R = 0.857; p = 0.010) (Figure 6). or even a total enhancement (26% improved, and there were no differences between post-condition and control group). These DISCUSSION two degrees of positive changes were also found in scores related to global cognitive functioning such as Full-Scale IQ, The present study aimed to provide evidence of the PIQ, Speed Processing Index, Working Memory Index, or POI. neurophysiological mechanisms underlying cognitive deficits In addition, some indicators related to functional performance, and changes in brain function associated with the recovery of such as DEX or PCRS (completed by relatives of patients), cognitive processes in stroke patients who underwent a NR. also improved after rehabilitation. This trend could represent Frontiers in Neurology | www.frontiersin.org 7 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke FIGURE 4 | (A) Significant FC results (p < 0.005, corrected) in the beta band (12–30 Hz) when comparing within the stroke patients’ group, the pre-condition and the post- condition. Line thickness of significant links is proportional to FC values (a higher value corresponds to thicker lines, and vice versa); (B) Significant functional connectivity links in the beta band are represented as bar graphs. Red color represents higher connectivity values for pre-condition compared to post-condition and blue color illustrates lower connectivity values for pre-condition compared to post-condition. ROIs included: lMFG, Left Middle Frontal Gyrus; rMFG, Right Middle Frontal Gyrus; lIFG, Left Inferior Frontal Parstriangularis; rIFG, Right Inferior Frontal Parstriangularis; rmOrbG, Right Medial Orbito Frontal Gyrus; lACCr, Left Rostral Anterior Cingulate; rACCr, Right Rostral Anterior Cingulate; lOrbG, Left Lateral Orbito Frontal Gyrus; rIFGorb, Right Inferior Frontal Orbital; lIFGop, Left Inferior Frontal Gyrus Opercular; lSTG, Left Superior Temporal Gyrus; rPreCG, Right Precentral Gyrus; rSTG, Right Superior Temporal Gyrus; lMTG, Left Middle Temporal Gyrus; rPoCG, Right Poscentral Gyrus; lPHG, Left Parahipocampal Gyrus; rSMG, Right Supramarginal Gyrus; lSMG, Left Supramarginal Gyrus; lPCUN, Left Precuneus; lstroke patientsL, Left Superior Parietal Lobule; rstroke patientsL, Right Superior Parietal Lobule; rLING, Right Lingual Cortex; rCAL, Right Calcarine; lMOG, Left Lateral Occipital Gyrus; rMOG, Right Lateral Occipital Gyrus. a partial cognitive and functional improvement and may have Based on these data, we can affirm that these stroke patients clinically relevant implications, since it may be considered as experienced at least some recovery in their global cognitive an indicator of recovery. To rule out the possibility that the capacity, despite the different etiology and location of their learning effect could be influencing the improvement of some lesions. Nonetheless the previous literature about the effect of cognitive scores, we have other complementary data related to NR on specific cognitive domains remains unclear. Low to changes in brain function. Specifically, the stroke patients of this moderate effects of rehabilitation in executive functions (6), study exhibited a widespread increased FC pattern within the attention (4, 52), or memory (53) have been reported. These beta band, indicating that their original disruption was restored results could have low consistency for different reasons: (1) in the recording performed after NR in that frequency band. the low methodological quality or insufficient description (2, 4, We also obtained two very important results associated with the 52) the use of small samples; (3) the absence of comparisons relationships between cognitive scores and changes in FC. On between intervention and no intervention or placebo conditions the one hand, we found three positive and significant recovery (4, 6) the deficit of randomized control trials (4, 5, 52) the signatures correlations between the FC strength ratio and three need for standardized definition and outcome measures (53, 54); cognitive measures changes (in Full-Scale IQ, BNT, and LM1), and (6) the lack of inclusion of functional ability measures in and on the other hand, we observed the predictive capacity of the rehabilitation outcome evaluation (52). It seems important some neuropsychological test scores (in the pre-condition) and to find the most effective procedures to try recovering the the recovery of the brain network (in the FC strength ratio). In cognitive deficits associated with stroke, considering that the this sense, we found two predictive markers of brain recovery prevalence of post-stroke cognitive impairment is 53.4% (55), related to two global cognitive domains, PIQ and Perceptual since it causes an increase in the institutionalization rate and Organization (both from the WAIS-III scale). costs of care (56, 57) and a decrease in the quality of life (58). Frontiers in Neurology | www.frontiersin.org 8 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke non-invasive transcranial magnetic stimulation to enhance some cognitive recoveries in stroke patients (60). Restoration interventions aim to regain the cognitive abilities of stroke patients, including domain-specific interventions and treatments for generalized cognitive impairment (10). The patients of our study received an integrated treatment based on the holistic-comprehensive model proposed by Ben-Yishay and Diller (41), which is consistent with the interventions suggested by some experts in post-stroke cognitive rehabilitation in terms of their global treatment approach. This type of treatment could be very successful for this type of patients, insofar as it produces more clearly a pattern of overall improvement, both at behavioral and brain level. Other types of cognitive interventions, such as computer-assisted cognitive rehabilitation that has increased in recent years, although show some efficacy in improving attention, memory, executive function, or visuo-spatial neglect in stroke patients (61, 62), present very limited effects on working memory and even no effects on cognitive function compared to healthy controls (63). As discussed above, overall review studies on the effectiveness of cognitive intervention with stroke patients do not provide clear conclusions. However, we must know that one of the most important issues regarding the functioning of the human mind has to do with the factor of interdependence between the different cognitive domains. This aspect is often overlooked in cognitive performance studies, and review studies of the effectiveness of cognitive treatments do not usually consider it. For example, there are studies that focus on improving attention after having specifically trained it, and thus with the rest of cognitive domains, without evaluating the impact of attention deficit or executive deficit in other domains such as memory or language. However, cognitive interdependence makes it very exceptional for patients who have brain injuries to suffer a specific cognitive deficit in a specific cognitive domain. The cognitive deficit of brain injury patients usually affects several domains, for example, visuo-spatial attention, working memory, executive functions, and episodic memory. Thus, trying to understand functioning of human cognition from independent cognitive domains, is probably an incorrect approach that hinders the interpretation of FIGURE 5 | Recovery signatures correlations. Statistically significant correlations between cognitive differences and the mean beta strength ratio in the results in neuropsychology. Usual intervention in the clinical the stroke patients’ group. (A) Full Scale IQ (of WAIS-III); (B) BNT (Boston setting is not as domain specific as studies suggest, since isolating Naming Test); (C) WMS-LM1 (Logical Memory 1 of WMS-III). cognitive processes in habitual actions is not easy. However, the neuropsychological literature continues to try to understand the effect of rehabilitation on each cognitive domain individually. This discrepancy requires a revision and a paradigm shift. In addition, if stroke is a central factor in the development of Furthermore, our intention was to go one step further cognitive impairment, or if this depends on the severity, subtype, trying to understand whether this recovery process seen at a location, or its recurrence, it becomes essential to understand the cognitive and behavioral level could have some reflection in brain mechanisms that produce both deficits and their recovery. brain functioning. Cognitive functions depend on the integrated There is agreement around the idea that cognitive rehabilitation functioning of large-scale distributed brain networks (64). interventions aim to improve the impaired brain functions in Specifically, recent evidence suggests that FC between brain stroke patients, and that it must be related to the damaged regions may play an important role when difficulties arise from anatomical substrate (10). Usually, rehabilitation facilitates the deficits in attention, memory, or other cognitive functions (65). development of behavioral and cognitive strategies that have a In this context, we firstly looked for a FC pattern related to stroke. positive impact on the structural and functional recovery of the We found a disruption in the pattern of brain functioning, with brain (53, 59). In this sense, it seems worthy to have in mind other a significant decrease in the beta band FC for intra and inter- promising complementary interventions including for example hemispheric connections in stroke patients before rehabilitation Frontiers in Neurology | www.frontiersin.org 9 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke FIGURE 6 | Recovery prediction. Statistically significant differences between the cognitive scores in the pre-condition and the mean beta strength ratio (post/pre) in the stroke patients’ group. (A) Performance IQ (WAIS-III); (B) POI (Perceptual Organization Index of WAIS-III); (C) Block Design, Matrix Reasoning, and Picture Completion (of WAIS-III). Note that the tests represented in this figure are related to each other, since Block Design, Matrix Reasoning, and Picture Completion are the tests included in Perceptual Organization Index, and POI (Perceptual Organization Index) is included in Performance IQ (on the WAIS-III scale). compared to healthy controls. General disruption of dynamic Previous results have described the role of RS FC as a predictor networks after stroke have been previously reported in MEG of motor learning ability in beta-band for healthy participants studies (25). Alterations in beta band activity have been especially (74) or as a predictor of post-stroke motor recovery in alpha- related to stroke compared to healthy controls (26), supporting band (75). Furthermore, the reorganization in FC in the beta- our results. band during resting-state has previously been associated with the On the other hand, understanding the interaction between success of cognitive and physical interventions (13, 76, 77). brain regions within a network (i.e., their FC), and the Up to this point, two independent markers of functional interactions among networks are both important for efficient recovery (i.e., cognitive, and neurophysiological) were found cognitive function (66). Therefore, exploring the possible in our stroke patients who went through the NR, but we neurophysiological mechanisms underlying the recovery process developed further analyses to discover the possible relation after stroke seems to be a crucial point in understanding the between them. As mentioned before, trying to understand effect of cognitive rehabilitation on the brain. By observing complex systems such as human cognition or brain functioning stroke patients before and after the rehabilitation, a specific focusing on only some of their components gives partial brain recovery pattern emerged, characterized by a widespread information of the entire process. Therefore, with the aim increased FC in the beta in the post-condition. The functioning to simplify the dimensionality of the data and to explore of the beta frequency band has previously been related, in relationships between cognitive and neurophysiological findings, healthy population, to different cognitive tasks such as working the difference (D = Post–Pre) of the most representative memory (67–69), attention (70), and motor performance (71). cognitive scores and the ratio of change (post/pre) of the total On the contrary, our neurophysiological data were acquired FC beta network strength were used. Three positive correlations during resting state (RS) which has been shown that is the most between recovery signatures (i.e., cognitive and neurophysiology) stable condition across patients with different symptomatology were observed for stroke patients, corresponding to Full and it also has been considered a hallmark for clinical diagnosis Scale IQ of WAIS-III, Boston Naming Test, and Logical and monitoring the recovery of patients that underwent a Memory 1 of WMS-III. These results showed that the beta rehabilitation, both in MRI (18, 72) and MEG studies (22, 73). connectivity changes after the NR, compared with the data Frontiers in Neurology | www.frontiersin.org 10 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke obtained in the first recording are, in fact, the reflection of LIMITATIONS AND FUTURE RESEARCH the cognitive improvement in the brain. The measures related with improvement represent global neuropsychological indices, An important contribution of this study may be that, unlike which contain different cognitive domains such as sustained most studies, functional brain connectivity was measured with and switching attention, visuo-spatial attention, visuo-spatial MEG. Despite the intrinsic limitations of BOLD fMRI, MEG working memory, planning, flexibility, or processing speed, is a measure of brain activity with incredibly high temporal but also episodic memory or verbal denomination. Initially, resolution (ms). Despite its advantages, MEG is an underused the neuropsychological deficit observed in stroke patients was neuroimaging tool in clinical and research contexts. Although global and their subsequent cognitive recovery, although not the sample of this exploratory study size was small, we were able complete, was also general. The scope of cognitive changes to identify a pattern of recovery of FC in the beta band related after NR was really wide, probably due to the type of to cognitive enhancement in stroke patients who underwent a intervention, which was holistic and not only focused on NR. Additionally, we were able to make predictions based on specific cognitive functions, including global and interdependent the cognitive performance of stroke patients before rehabilitation domains, and focusing on individual cognitive, functional, about the future functional restoration of the brain network. emotional, and behavioral imbalances. All these evidences are Thus, larger MEG studies with stroke patients are needed to consistent with the brain network global changes observed in demonstrate the power of this neurophysiological tool within stroke patients. this neurological field. Another limitation that this study faces While addressing brain and cognitive changes in stroke is the heterogeneity of stroke patients, in terms of etiology and patients is important to understand the underlying mechanisms location of the injury. However, we believe that this heterogeneity of stroke and brain plasticity, the early detection of patterns has provided an interesting approach to the study since it or biomarkers is also relevant to predict which subjects are has allowed us to explore the effect of cognitive intervention more likely to improve and benefit from neurorehabilitation. both at cognitive and neurophysiological level in stroke patients In this regard, adjustments can be made for those patients with different etiology. Furthermore, another limitation of the who will not benefit from this option. Thus, we performed study is the absence of a clinical control group (i.e., stroke an additional correlation analysis in which the pre-condition patients without rehabilitation). Considering this limitation, we cognitive measures and the ratio of change (post/pre) of the cannot assure that cognitive improvement is due specifically total FC beta network strength were taken into account. Two or uniquely to cognitive rehabilitation because it could also global cognitive markers were stated as predictors of brain represent some degree of spontaneous clinical recovery after functioning recovery, PIQ and POI. Furthermore, within POI, stroke. In accordance with the aforementioned obstacles, future we found a significant association for every subtest included in studies should include a group of stroke patients without the global index (i.e., Picture Completion, Matrix Reasoning, cognitive intervention (e.g., on the waiting list), and larger and Block Design). In the clinical setting, the role of prediction samples of patients that allow comparisons based on the etiology and the location of the lesion. Finally, it would be interesting to in terms of the degree of future recovery has always been focus on different networks of the brain (78) such as the default important, however it is a complex and complicated issue. mode, the salience, or the executive control networks (79), to Until now we only had clinical, cognitive, behavioral, and explore specific changes in each network recovery pattern, and social variables, but these results indicate that the relationship its possible relationship to particular improvements in cognition between behavior and the brain can contribute to this topic. after stroke. In this case, the results indicate that the initial state around some cognitive domains such as visuo-spatial attention, visuo- spatial working memory, or planning capacity, could have DATA AVAILABILITY STATEMENT a very relevant role in the evolution and recovery of the The raw data supporting the conclusions of this article will be brain network of stroke patients. This information is not only made available by the authors, without undue reservation. important to predict the patients who will improve the most, but it can also serve to think about more powerful intervention ETHICS STATEMENT procedures for those patients who have a more serious deficit around these cognitive domains. This result has very relevant The studies involving human participants were reviewed and clinical implications. approved by the National Brain Injury Rehabilitation Center In conclusion, if NR aims to improve people’s cognitive Ethics Committee (Madrid). The patients/participants provided function in order to restore their general performance and their written informed consent to participate in this study. independence in functional activities, the results of the present study are in line with this objective, showing a clear AUTHOR CONTRIBUTIONS improvement pattern in stroke patients who received NR both cognitively and brain function. We are also sure that this SP, LT-S, ML, NP, and FM designed research. SP and LT-S study points out the importance of including neuropsychological performed main calculations of the study and prepared figures. and neurophysiological variables in the assessment of the SP, LT-S, ML, BC, LC, AB, FM, and NP collaborated actively in outcome and effectiveness of psychological interventions of writing the manuscript. All authors contributed to the article and stroke patients. approved the submitted version. Frontiers in Neurology | www.frontiersin.org 11 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke FUNDING and Technological Innovation (CONCYTEC, Perú) through the National Fund for Scientific and Technological Development Financial support of the project was provided by IMSERSO (07- (FONDECYT, Perú) to BC. 2008) and the Spanish MICINN (PSI2011-28388). Research by SP was supported by the Spanish MINECO post-doctoral fellowship SUPPLEMENTARY MATERIAL (FJC2019-041205-I). Additionally, this work was supported by a predoctoral researcher grant from Universidad Complutense de The Supplementary Material for this article can be found Madrid (CT42/18-CT43/18) and co-founded by Santander Bank online at: https://www.frontiersin.org/articles/10.3389/fneur. to LT-S, and by the National Council of Science, Technology 2022.838170/full#supplementary-material REFERENCES networks after stroke. J Cereb Blood Flow Metab. (2013) 33:1279– 85. doi: 10.1038/jcbfm.2013.80 1. Feigin VL, Norrving B, Mensah GA. Global burden of stroke. Circ Res. (2017) 16. Bayrak S, Khalil AA, Villringer K, Fiebach JB, Villringer A, Margulies DS, et 120:439–48. doi: 10.1161/CIRCRESAHA.116.308413 al. The impact of ischemic stroke on connectivity gradients. NeuroImage Clin. 2. Aam S, Einstad MS, Munthe-Kaas R, Lydersen S, Ihle-Hansen H, Knapskog (2019) 24:101947. doi: 10.1016/j.nicl.2019.101947 AB, et al. Post-stroke cognitive impairment—impact of follow-up time and 17. Carrera E, Tononi G. Diaschisis: Past, present, future. Brain. (2014) 137:2408– stroke subtype on severity and cognitive profile: the Nor-COAST study. Front 22. doi: 10.1093/brain/awu101 Neurol. (2020) 11:699. doi: 10.3389/fneur.2020.00699 18. Baldassarre A, Ramsey LE, Siegel JS, Shulman GL, Corbetta M. Brain 3. Al-Qazzaz NK, Ali SH, Ahmad SA, Islam S, Mohamad K. Cognitive connectivity and neurological disorders after stroke. Curr Opin Neurol. (2016) impairment and memory dysfunction after a stroke diagnosis: a post- 29:706–13. doi: 10.1097/WCO.0000000000000396 stroke memory assessment. Neuropsychiatr Dis Treat. (2014) 10:1677– 19. Ramsey LE, Siegel JS, Baldassarre A, Metcalf NV, Zinn K, Shulman GL, et al. 91. doi: 10.2147/NDT.S67184 Normalization of network connectivity in hemispatial neglect recovery. Ann 4. Loetscher T, Potter KJ, Wong D, das Nair R. Cognitive rehabilitation for Neurol. (2016) 80:127–41. doi: 10.1002/ana.24690 attention deficits following stroke. Cochrane Database Syst Rev. (2019) 20. Maestú F, Peña JM, Garcés P, González S, Bajo R, Bagic A, et al. A multicenter 2019:CD002842. doi: 10.1002/14651858.CD002842.pub3 study of the early detection of synaptic dysfunction in mild cognitive 5. Feigin VL, Roth GA, Naghavi M, Parmar P, Krishnamurthi R, Chugh S, et al. impairment using magnetoencephalography-derived functional connectivity. Global burden of stroke and risk factors in 188 countries, during 1990–2013: Neuroimage Clin. (2015) 9:103–9. doi: 10.1016/j.nicl.2015.07.011 a systematic analysis for the Global Burden of Disease Study 2013. Lancet 21. Stam CJ. Use of magnetoencephalography (MEG) to study functional brain Neurol. (2016) 15:913–24. doi: 10.1016/S1474-4422(16)30073-4 networks in neurodegenerative disorders. J Neurol Sci. (2010) 289:128– 6. Chung CSY, Pollock A, Campbell T, Durward BR, Hagen S. Cognitive 34. doi: 10.1016/j.jns.2009.08.028 rehabilitation for executive dysfunction in adults with stroke or other adult 22. Paggiaro A, Birbaumer N, Cavinato M, Turco C, Formaggio E, del Felice A, non-progressive acquired brain damage. Cochrane Database Syst Rev. (2013) et al. Magnetoencephalography in stroke recovery and rehabilitation. Front 2013:CD008391. doi: 10.1002/14651858.CD008391.PUB2 Neurol. (2016) 7:35. doi: 10.3389/fneur.2016.00035 7. Rasquin SMC, Lodder J, Ponds RWHM, Winkens I, Jolles J, Verhey FRJ. 23. Tewarie P, Steenwijk MD, Tijms BM, Daams M, Balk LJ, Stam CJ, et al. Cognitive functioning after stroke: a one-year follow-up study. Dement Disruption of structural and functional networks in long-standing multiple Geriatr Cogn Disord. (2004) 18:138–44. doi: 10.1159/000079193 sclerosis. Hum Brain Mapp. (2014) 35:5946–61. doi: 10.1002/HBM.22596 8. Katz N, Hartman-Maeir A, Ring H, Soroker N. Relationships of cognitive 24. Pusil S, López ME, Cuesta P, Bruña R, Pereda E, Maestú F. performance and daily function of clients following right hemisphere stroke: Hypersynchronization in mild cognitive impairment: the ‘X’ model. Brain. predictive and ecological validity of the LOTCA battery. Occup Ther J Res. (2019) 142:3936–50. doi: 10.1093/brain/awz320 (2016) 20:3–17. doi: 10.1177/153944920002000101 25. Marsh EB, Brodbeck C, Llinas RH, Mallick D, Kulasingham JP, Simon JZ, et 9. Ferreira MGR, Moro CHC, Franco SC. Desempenho cognitivo após al. Poststroke acute dysexecutive syndrome, a disorder resulting from minor acidente vascular cerebral isquêmico. Dement Neuropsychol. (2015) 9:165– stroke due to disruption of network dynamics. Proc Natl Acad Sci USA. (2020) 75. doi: 10.1590/1980-57642015DN92000011 117:33578–585. doi: 10.1073/pnas.2013231117/ 10. Zhao Q, Wang X, Wang T, Dmytriw AA, Zhang X, Yang K, et al. 26. Kulasingham JP, Brodbeck C, Khan S, Marsh EB, Simon JZ. Bilaterally Cognitive rehabilitation interventions after stroke: protocol for a systematic reduced rolandic beta band activity in minor stroke patients. bioRxiv Preprint. review and meta-analysis of randomized controlled trials. Syst Rev. (2021) (2021). doi: 10.1101/2021.10.15.464457 10:66. doi: 10.1186/s13643-021-01607-7 27. Castellanos NP, Leyva I, Buldú JM, Bajo R, Paúl N, Cuesta P, et al. Principles 11. Canuet L. Neurorehabilitation in stroke: the role of functional connectivity. of recovery from traumatic brain injury: reorganization of functional Int J Neurorehabil. (2015) 2:172. doi: 10.4172/2376-0281.1000172 networks. Neuroimage. (2011) 55:1189–99. doi: 10.1016/j.neuroimage.201 12. Cicerone KD, Mott T, Azulay J, Sharlow-Galella MA, Ellmo WJ, Paradise S, 0.12.046 et al. Randomized controlled trial of holistic neuropsychologic rehabilitation 28. Teasdale G, Jennett B. Assessment of coma and impaired consciousness a after traumatic brain injury. Arch Phys Med Rehabil. (2008) 89:2239– practical scale. Lancet. (1974) 2:81–4. doi: 10.1016/s0140-6736(74)91639-0 49. doi: 10.1016/j.apmr.2008.06.017 29. Wechsler D. Wechsler Adult Intelligence Scale (WAIS-III). San Antonio, TX: 13. Castellanos NP, Paúl N, Ordóñ Ez VE, Demuynck O, Bajo R, Campo Psychological Corporation (1997). P, et al. Reorganization of functional connectivity as a correlate of 30. Schretlen D. Brief Test of Attention. Baltimore, MD: Psychological Assessment cognitive recovery in acquired brain injury. Brain. (2010) 133:2365– Resources (1989). 81. doi: 10.1093/brain/awq174 31. Reitan RM, Wolfson D. The Halstead-Reitan Neuropsychological Test 14. Alexander LD, Black SE, Gao F, Szilagyi G, Danells CJ, McIlroy WE. Battery e Theory and Clinical Interpretation. Tucson, AZ: Neuropsychology Correlating lesion size and location to deficits after ischemic stroke: the Press (1993) influence of accounting for altered peri-necrotic tissue and incidental silent 32. Golden CJ. A Manual for the Clinical and Experimental Use of the Stroop Color infarcts. Behav Brain Funct. (2010) 6:6. doi: 10.1186/1744-9081-6-6 and Word Test. Chicago, IL: Stoelting (1978). 15. Ovadia-Caro S, Villringer K, Fiebach J, Jungehulsing GJ, van der Meer 33. Jensen AR, Rohwer WD. The Stroop color-word test: a review. Acta Psychol. E, Margulies DS, et al. Longitudinal effects of lesions on functional (1966) 25:36–93. doi: 10.1016/0001-6918(66)90004-7 Frontiers in Neurology | www.frontiersin.org 12 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke 34. Heaton RK, Grant I, Matthews CG. Comprehensive Norms for an Expanded 53. das Nair R, Cogger H, Worthington E, Lincoln NB. Cognitive rehabilitation Halstead-Reitan Battery: Demographic Corrections, Research Findings, and for memory deficits after stroke. Cochrane Database Syst Rev. (2016) Clinical Applications. Odessa, FL: Psychological Assessment Resources. (1991). 9:CD002293. doi: 10.1002/14651858.CD002293.pub3 35. Édouard L. Récréations Mathématiques. Vol III. Paris: Reprinted 54. Bernhardt J, Hayward KS, Kwakkel G, Ward NS, Wolf SL, Borschmann K, et several times by Albert Blanchard (1983). Available online at: https:// al. Agreed definitions and a shared vision for new standards in stroke recovery books.google.es/books?hl=en&lr=&id=hwYAAAAAQAAJ&oi=fnd& research: the Stroke Recovery and Rehabilitation Roundtable taskforce. Int J pg=PA3&dq=%C3%89douard$+$L.$+$R%C3%A9cr%C3%A9ations$+ Stroke. (2017) 12:444–50. doi: 10.1177/1747493017711816 $Math%C3%A9matiques.$+$vol.$+$III&ots=6q1xOXhLC4&sig= 55. Barbay M, Diouf M, Roussel M, Godefroy O. Systematic review and bZgxKMrYvQEczRsMbW7m9ovm5WM&redir_esc=y#v=onepage&q= meta-analysis of prevalence in post-stroke neurocognitive disorders in %C3%89douard%20L.%20R%C3%A9cr%C3%A9ations%20Math%C3 hospital-based studies. Dement Geriatr Cogn Disord. (2019) 46:322– %A9matiques.%20vol.%20III&f=false (accessed December 15, 2021). 34. doi: 10.1159/000492920 36. Wilson BA, Alderman N, Burgess PW, Emslie H, Evans JJ. BADS: Behavioural 56. Pasquini M, Leys D, Rousseaux M, Pasquier F, Hénon H. Influence of cognitive Assessment of the Dysexecutive Syndrome : Test Manual. London: Thames impairment on the institutionalisation rate 3 years after a stroke. J Neurol Valley Test Company (1996). Neurosurg Psychiatry. (2007) 78:56–9. doi: 10.1136/jnnp.2006.102533 37. Kaplan E. Boston Naming Test. Philadelphia, PA: Lea & Febiger (1983). 57. Claesson L, Lindén T, Skoog I, Blomstrand C. Cognitive impairment after 38. Benton AL, Hamsher K. Controlled Oral Word Association Test – COWAT. stroke - Impact on activities of daily living and costs of care for elderly Iowa City, IA: University of Iowa Press (1976) people: the Göteborg 70+ stroke study. Cerebrovasc Dis. (2005) 19:102– 39. Sedó M. FDT. Test de los Cinco Dígitos. TEA Ediciones (2007). Available 9. doi: 10.1159/000082787 online at: https://web.teaediciones.com/fdt--test-de-los-cinco-digitos.aspx 58. Park JH, Kim BJ, Bae H-J, Lee J, Lee J, Han M-K, et al. Impact of post-stroke (accessed December 15, 2021). cognitive impairment with no dementia on health-related quality of life. J 40. Prigatano GP, Schacter DL. Awareness of Deficit After Brain Injury. New York, Stroke. (2013) 15:49–56. doi: 10.5853/JOS.2013.15.1.49 NY: Oxford University Press (1991). 59. Robertson S. The efficacy of oro-facial and articulation exercises in 41. Ben-Yishay Y, Diller L. Handbook of Holistic Neuropsychological dysarthria following stroke Int J Lang Commun Disord. (2001) 36(Suppl):292– Rehabilitation: Outpatient Rehabilitation of Traumatic Brain Injury. 7. doi: 10.3109/13682820109177900 Oxford University Press (2011). Available online at: https://books. 60. Draaisma LR, Wessel MJ, Hummel FC. Non-invasive brain stimulation google.es/books?hl=es&lr=&id=Yzg_zb94ZHcC&oi=fnd&pg=PP1&dq= to enhance cognitive rehabilitation after stroke. Neurosci Lett. (2020) Handbook$+$of$+$Holistic$+$Neuropsychological$+$Rehabilitation:$+ 719:133678. doi: 10.1016/j.neulet.2018.06.047 $Outpatient$+$Rehabilitation$+$of$+$Traumatic$+$Brain$+$Injury&ots= 61. Yoo C, Yong MH, Chung J, Yang Y. Effect of computerized cognitive KzE1fPTgBK&sig=0neZLIlQmPk85DCW3BVyrj5qVPQ#v=onepage&q= rehabilitation program on cognitive function and activities of living in stroke Handbook%20of%20Holistic%20Neuropsychological%20Rehabilitation%3A patients. J Phys Ther Sci. (2015) 27:2487–9. doi: 10.1589/jpts.27.2487 %20Outpatient%20Rehabilitation%20of%20Traumatic%20Brain%20Injury& 62. Svaerke K, Niemeijer M, Mogensen J, Christensen H. The effects of f=false (accessed December 15, 2021). computer-based cognitive rehabilitation in patients with visuospatial neglect 42. Bell AJ, Sejnowski TJ. An information-maximization approach to blind following stroke: a systematic review. Top Stroke Rehabil. (2019) 26:214– separation and blind deconvolution. Neural Comput. (1995) 7:1129– 25. doi: 10.1080/10749357.2018.1556963 59. doi: 10.1162/NECO.1995.7.6.1129 63. Wentink MM, Berger MAM, de Kloet AJ, Meesters J, Band GPH, Wolterbeek 43. Tadel F, Baillet S, Mosher JC, Pantazis D, Leahy RM. Brainstorm: a user- R, et al. The effects of an 8-week computer-based brain training programme friendly application for MEG/EEG analysis. Comput Intell Neurosci. (2011) on cognitive functioning, QoL and self-efficacy after stroke. Neuropsychol 2011:879716. doi: 10.1155/2011/879716 Rehabil. (2016) 26:847–65. doi: 10.1080/09602011.2016.1162175 44. Holmes CJ, Hoge R, Collins L, Woods R, Toga AW, Evans AC. 64. Mesulam MM. From sensation to cognition. Brain. (1998) 121(Pt 6):1013– Enhancement of MR images using registration for signal averaging. J 52. doi: 10.1093/brain/121.6.1013 Comput Assist Tomogr. (1998) 22:324–33. doi: 10.1097/00004728-19980300 65. Siegel JS, Ramsey LE, Snyder AZ, Metcalf NV, Chacko RV, Weinberger K, 0-00032 et al. Disruptions of network connectivity predict impairment in multiple 45. Mosher JC, Baillet S, Leahy RM. Equivalence of linear approaches behavioral domains after stroke. Proc Natl Acad Sci USA. (2016) 113:E4367– in bioelectromagnetic inverse solutions. In: IEEE Workshop on 76. doi: 10.1073/pnas.1521083113 Statistical Signal Processing Proceedings. St. Louis, MO (2003) p. 66. Leech R, Kamourieh S, Beckmann CF, Sharp DJ. Fractionating the 294–7. doi: 10.1109/SSP.2003.1289402 default mode network: distinct contributions of the ventral and dorsal 46. Palva S, Palva JM. Discovering oscillatory interaction networks with posterior cingulate cortex to cognitive control. J Neurosci. (2011) 31:3217– M/EEG: challenges and breakthroughs. Trends Cogn Sci. (2012) 16:219– 24. doi: 10.1523/JNEUROSCI.5626-10.2011 30. doi: 10.1016/j.tics.2012.02.004 67. Koshy SM, Wiesman AI, Proskovec AL, Embury CM, Schantell MD, Eastman 47. Hassan M, Dufor O, Merlet I, Berrou C, Wendling F. EEG source connectivity JA, et al. Numerical working memory alters alpha-beta oscillations and analysis: from dense array recordings to brain networks. PLoS ONE. (2014) connectivity in the parietal cortices. Hum Brain Mapp. (2020) 41:3709– 9:e105041. doi: 10.1371/JOURNAL.PONE.0105041 19. doi: 10.1002/hbm.25043 48. Desikan RS, Ségonne F, Fischl B, Quinn BT, Dickerson BC, Blacker D, et al. 68. Proskovec AL, Heinrichs-Graham E, Wilson TW. Load modulates the alpha An automated labeling system for subdividing the human cerebral cortex on and beta oscillatory dynamics serving verbal working memory. Neuroimage. MRI scans into gyral based regions of interest. Neuroimage. (2006) 31:968– (2019) 184:256–65. doi: 10.1016/j.neuroimage.2018.09.022 80. doi: 10.1016/j.neuroimage.2006.01.021 69. Schmidt R, Ruiz MH, Kilavik BE, Lundqvist M, Starr PA, Aron 49. Mormann F, Lehnertz K, David P, Elger CE. Mean phase coherence AR. Beta oscillations in working memory, executive control of as a measure for phase synchronization and its application to the movement and thought, and sensorimotor function. J Neurosci. EEG of epilepsy patients. Phys D Nonlin Phenom. (2000) 144:358– 39:8231–8. doi: 10.1523/JNEUROSCI.1163-19.2019 69. doi: 10.1016/S0167-2789(00)00087-7 70. Kamiński J, Brzezicka A, Gola M, Wróbel A. Beta band oscillations 50. Bruña R, Maestú F, Pereda E. Phase locking value revisited: teaching new tricks engagement in human alertness process. Int J Psychophysiol. (2012) 85:125– to an old dog. J Neural Eng. (2018) 15:056011. doi: 10.1088/1741-2552/aacfe4 8. doi: 10.1016/j.ijpsycho.2011.11.006 51. Benjamini Y, Yekutieli D. The control of the false discovery rate 71. Chung JW, Ofori E, Misra G, Hess CW, Vaillancourt DE. Beta- in multiple testing under dependency. Ann Statist. (2001) 29:1165– band activity and connectivity in sensorimotor and parietal cortex are 88. doi: 10.1214/aos/1013699998 important for accurate motor performance. Neuroimage. (2017) 144:164– 52. Hazelton C. Can cognitive rehabilitation improve attention deficits 73. doi: 10.1016/j.neuroimage.2016.10.008 following stroke? - A Cochrane Review summary with commentary 72. Dacosta-Aguayo R, Graña M, Savio A, Fernández-Andújar M, Millán M, NeuroRehabilitation. (2020) 47:355–7. doi: 10.3233/NRE-209007 López-Cancio E, et al. Prognostic value of changes in resting-state functional Frontiers in Neurology | www.frontiersin.org 13 February 2022 | Volume 13 | Article 838170 Pusil et al. Functional Recovery Network After Stroke connectivity patterns in cognitive recovery after stroke: a 3T fMRI pilot study. 79. Raichle ME, MacLeod AM, Snyder AZ, Powers WJ, Gusnard Hum Brain Mapp. (2014) 35:3819–31. doi: 10.1002/hbm.22439 DA, Shulman GL, et al. A default mode of brain function. 73. Martín-Buro MC, Garcés P, Maestú F. Test-retest reliability of resting-state Proc Nat Acad Sci USA. (2001) 98:676–82. doi: 10.1073/PNAS.9 magnetoencephalography power in sensor and source space. Hum Brain 8.2.676 Mapp. (2016) 37:179–90. doi: 10.1002/hbm.23027 74. Sugata H, Yagi K, Yazawa S, Nagase Y, Tsuruta K, Ikeda T, Conflict of Interest: The authors declare that the research was conducted in the et al. Role of beta-band resting-state functional connectivity absence of any commercial or financial relationships that could be construed as a as a predictor of motor learning ability. NeuroImage. (2020) potential conflict of interest. 210:116562. doi: 10.1016/j.neuroimage.2020.116562 75. Westlake KP, Hinkley LB, Bucci M, Guggisberg AG, Findlay AM, Henry RG, et Publisher’s Note: All claims expressed in this article are solely those of the authors al. Resting state alpha-band functional connectivity and recovery after stroke. Exp Neurol. (2012) 237:160–9. doi: 10.1016/j.expneurol.2012.06.020 and do not necessarily represent those of their affiliated organizations, or those of 76. Weiss SA, Bassett DS, Rubinstein D, Holroyd T, Apud J, Dickinson D, et al. the publisher, the editors and the reviewers. Any product that may be evaluated in Functional brain network characterization and adaptivity during task practice this article, or claim that may be made by its manufacturer, is not guaranteed or in healthy volunteers and people with schizophrenia. Front Hum Neurosci. endorsed by the publisher. (2011) 5:81. doi: 10.3389/fnhum.2011.00081 77. Klados MA, Styliadis C, Frantzidis CA, Paraskevopoulos E, Bamidis PD. Beta- Copyright © 2022 Pusil, Torres-Simon, Chino, López, Canuet, Bilbao, Maestú and band functional connectivity is reorganized in mild cognitive impairment Paúl. This is an open-access article distributed under the terms of the Creative after combined computerized physical and cognitive training. Front Neurosci. Commons Attribution License (CC BY). The use, distribution or reproduction in (2016) 10:55. doi: 10.3389/fnins.2016.00055 other forums is permitted, provided the original author(s) and the copyright owner(s) 78. Power JD, Cohen AL, Nelson SM, Wig GS, Barnes KA, Church JA, et al. are credited and that the original publication in this journal is cited, in accordance Functional network organization of the human brain. Neuron. (2011) 72:665– with accepted academic practice. No use, distribution or reproduction is permitted 78. doi: 10.1016/j.neuron.2011.09.006 which does not comply with these terms. Frontiers in Neurology | www.frontiersin.org 14 February 2022 | Volume 13 | Article 838170