Yuping Huang1†
Xiaoxuan Zhan2†Huizi Zeng2Shuyin Li1Jingqin Shi2Zhenhua Cui2Qianqian Fan2Binbin Li2Yanfang Sui2Fengyan Liang2,3*Zhenhua Song2*- 1School of Rehabilitation Medicine, Shandong University of Traditional Chinese Medicine, Jinan, China
- 2Department of Rehabilitation Medicine, Haikou Hospital Affiliated to Xiangya Medical College of Central South University, Haikou, China
- 3School of Biomedical Engineering, Hainan University, Sanya, China
Precision in assessing neurological function after stroke is key to optimizing the efficacy of rehabilitation. Functional near-infrared spectroscopy (fNIRS) provides a highly ecologically valid assessment of cortical activation and functional reorganization after stroke by monitoring cortical hemodynamic changes during different tasks. However, the current fNIRS task paradigm lacks systematic integration for standardized design and clinical translation strategies, and fragmented evidence is difficult to converge into actionable practice guidelines. To fill this gap, this paper systematically reviews the application of fNIRS in motor, cognitive, language, and dual-task paradigms in stroke rehabilitation research. It reveals the clinical value of different paradigms for neurological function assessment and proposes adaptive task designs that fit the functional characteristics of patients with stroke. This study emphasizes the importance of personalized and ecological paradigms, providing a theoretical basis and practical reference for subsequent standardized research on fNIRS task paradigms and developing clinical application standards.
1 Introduction
Stroke, the second most common fatal disease worldwide, accounts for approximately 11.6 million new cases and 5.5 million deaths annually, with more than 50% of survivors left with long-term functional impairment (GBD 2019 Stroke, 2021). The core goal of stroke rehabilitation is to promote functional recovery through neuroplastic remodeling (Singh, 2024), and accurate assessment of functional reorganization of the brain is key to optimizing rehabilitation strategies and achieving individualized interventions. Traditional clinical scales are highly subjective and insufficiently sensitive, creating challenges in capturing early neurological changes; functional magnetic resonance imaging (fMRI), although with high spatial resolution, is environmentally restricted and not applicable to natural rehabilitation monitoring (Chen et al., 2020). In contrast, functional near-infrared spectroscopy (fNIRS) with its portability, resistance to motion artifacts, and high temporal resolution is ideal for stroke rehabilitation assessment (Wang et al., 2025b).
Several stroke rehabilitation studies have used fNIRS. In the field of limb function rehabilitation, fNIRS is shown to detect oxygenated hemoglobin (HbO) concentration in motor networks represented by the sensorimotor cortex (SMC), supplementary motor area (SMA), premotor cortex (PMC), and prefrontal cortex (PFC) during tasks, such as motor imagery (Wang et al., 2022), robotic-assisted training (Li H. et al., 2023; Liu P. et al., 2024), standing balance (Xia Y. et al., 2022), and walking (Lim et al., 2022b). Lee S. H. et al. (2020) observed a balanced cortical activation pattern (decreased activation in SMC and SMA) in patients during the later stages of robot-assisted walking training. In cognitive rehabilitation, functional connectivity decline is considered a promising evaluative indicator for recognizing cognitive dysfunction after stroke (Ai et al., 2025; Zou et al., 2023). Moreover, fNIRS shows that patients with stroke cognitive impairment depend on compensatory activation in the right prefrontal lobe during working memory tasks (Ai et al., 2025). In aphasia, an fNIRS study showed that different interventions could have different pathways to recovery. Head acupuncture combined with speech training significantly increased HbO levels in the left frontal pole region to improve some naming functions (Lin et al., 2025); low-frequency repetitive transcranial magnetic stimulation (rTMS) improved naming performance by decreasing activation in multiple functional areas of language, such as the left superior temporal gyrus (STG), Broca's area, and others (Gan et al., 2024). Together, these studies suggest that fNIRS is uniquely valuable in revealing post-stroke neuroplasticity and optimizing rehabilitation strategies.
Paradigms are crucial to neurofunctional imaging studies (such as fNIRS) to induce the activation of specific brain regions. They include core elements, such as task types, presentation methods, and time parameters (Cao et al., 2021). The scientific design of the task paradigm directly determines the detection ability of neuroimaging tools and their data quality (Zhang et al., 2022), acting as a key link connecting brain signal measurements with the interpretation of clinical significance. However, the significant heterogeneity of stroke patients in terms of motor ability, cognitive function, attention maintenance, and fatigue tolerance poses challenges to designing task paradigms. Standardized tasks may not be adaptable to patients with different levels of dysfunction, resulting in missing data or reduced signal quality; oversimplified tasks may lack sensitivity to detect subtle neurological changes (Zhao et al., 2023). Therefore, designing fNIRS task paradigms that can accommodate patients' functional limitations while maintaining scientific validity is the key issue in neurological function assessment in stroke. Currently, applying the fNIRS task paradigm in stroke rehabilitation faces three major challenges: (1) Designing ecologically valid tasks applicable to patients with different dysfunctions (e.g., alternative exercise programs for hemiplegics); (2) Optimizing the task to support multimodal data integration so that the fNIRS signal can be complementarily validated using electromyography, movement parameters, and other metrics to enhance precise assessment of neural-behavioral associations; (3) Establishing a standardized paradigm library to enhance cross-study comparability and data sharing.
This systematic review will systematically address the research progress of the fNIRS task paradigm in stroke rehabilitation—excluding resting-state measurements—from four aspects: objectives, design principles, classification, and future trends. It will analyze paradigm design strategies based on neuroplasticity mechanisms, adaptive task tuning methods for patients' functional characteristics, and balanced solutions for clinical standardization and individualization. By integrating current evidence and proposing innovative perspectives, it aims to provide systematic guidance to researchers and promote the scientific application of the fNIRS task paradigm in stroke rehabilitation.
2 Research methods
The search was completed on February 19, 2025, using PubMed, Web of Science, Embase, and Scopus databases. The subject terms and free terms were mainly obtained from the MeSH database and Emtree and referred to relevant meta-analyses. The search terms “stroke” and “functional near-infrared spectroscopy” ranged from 2019 to 2024 were identified, and the search formula was constructed according to the rules of each database. The literature screening process (Figure 1) followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and was based on the following criteria:
Figure 1. Flowchart of literature search and screening.
Inclusion criteria:
(1) Stroke patients;
(2) The use of the fNIRS task-state experimental paradigm;
(3) Access to full text and extraction of key information;
(4) Literature in English;
(5) At least one of the following comparisons must be included: stroke patients vs. healthy controls; affected vs. unaffected cerebral regions; pre- vs. post-intervention; or response contrasts across distinct task difficulty levels or task content;
(6) Experimental trials, including randomized controlled trials (RCTs) and non-randomized controlled trials (non-RCTs), and observational studies.
Exclusion criteria:
(1) Not in style—review (including systematic review and meta-analysis), editorial, study protocol, case report, patent, grant award, conference paper, book chapter, letter;
(2) Irrelevant to the study topic (e.g., non-stroke subjects, non-rehabilitation studies, not using fNIRS or collecting resting-state data only);
(3) Poorly assessed study quality;
(4) Unavailability of full text or missing key information.
For included studies, key information was extracted: this included the content of the experimental paradigm based on the fNIRS task, fNIRS metrics (e.g., ΔHbO), regions of interest (ROIs), metrics for statistical analyses, and tools or procedures used for data processing and analysis.
3 Results and discussion
This review included 65 papers (42 non-randomized studies and 23 randomized controlled trials). The fNIRS paradigms used in the research were based on motor (67.69%), cognitive (12.31%), language (4.62%), swallowing (7.69%), and dual-task (7.69%). Table 1 lists the task design and brain regions of interest in the included studies. All studies were evaluated for quality using the Cochrane Risk of Bias 2.0 tool (RoB 2.0) for RCTs (Page et al., 2021), and the Methodological Index for Non-Randomized Studies (MINORS) (Slim et al., 2003) for non-randomized studies. Despite the exclusion of high-risk studies during the screening process, the majority of RCTs still demonstrated a high risk of bias. The primary sources of bias were identified in the areas of blinding, data integrity, and allocation concealment. Among the 13 non-randomized studies without a control group, all were classified as moderate risk, with MINORS scores ranging from 10 to 13. The main sources of bias in these studies included insufficient blinding, high attrition rates, and a lack of sample size calculation. In the 29 non-randomized studies with control groups, 75.86% were categorized as moderate risk (total score between 15 and 19), 13.79% as low risk (total score ≥20), and 10.34% as high risk (total score < 15). The predominant sources of bias in these studies were high attrition rates, inadequate sample sizes, and baseline inequivalence between groups. These issues significantly affected the reliability and generalizability of the results. Overall, while the included studies met basic quality standards, improvements are needed in areas such as blinding, attrition management, and sample size calculation to enhance the quality and reliability of future research.
Table 1. Task design and brain regions for included studies.
3.1 The objective of the fNIRS task paradigm
The fNIRS task paradigm aims to activate key brain regions through specific task design to dynamically resolve post-stroke neuroplasticity and optimize the objective basis of neurological level for rehabilitation strategies. Compared with traditional resting-state functional imaging, fNIRS can realize the transition from static observation to dynamic functional analysis by inducing activation of brain regions through task settings and capturing functional reorganization and compensation in real time. For example, the motor imagery (MI) task mainly activates the PFC and SMC (Kotegawa et al., 2020; Yu Y. et al., 2024; Hurst and Boe, 2022), the motor execution task significantly activates the primary motor cortex (M1) (Jalalvandi et al., 2024), and the cognitive task highlights the involvement of the PFC and parietal lobe (Heiberg et al., 2023). Different task paradigms can map neural network surrogate pathways: simple repetitive tasks (e.g., ankle dorsiflexion) are prone to standardization but underestimate network reorganization (Liang et al., 2022); complex dual-tasks (e.g., walking combined with mental arithmetic) expose deficits in the allocation of resources to the PFC (Wang Q. et al., 2023); ecological tasks such as virtual reality (VR) tasks activate a wider range of cortical networks (Parker et al., 2024). By rationally designing tasks, fNIRS parses neuroplasticity multidimensionally and provides a basis for assessing intervention effects and guiding personalized rehabilitation, making it a valuable tool for dynamic neurofunctional mapping.
3.2 Basic principles of task paradigm design
3.2.1 Time parameter design
The fNIRS task paradigm is mainly divided into two categories: block design and event-related design. The block design is the preferred paradigm for stroke research due to its high signal-to-noise ratio, and it obtains stable HbO and deoxygenated hemoglobin (HbR) signals by focusing on the presentation of similar stimuli for 20–30 s (Luke et al., 2021). In contrast, event-related designs use short 2–5 s stimuli to capture transient responses (Defenderfer et al., 2017). Based on hemodynamic response characteristics, a standard cycle of a 20–30-s task period (up to the oxygenation plateau) with a 30–40-s baseline period is recommended (Herold et al., 2017, 2018). In older stroke patients or those with poor functional status, appropriately prolonged rest periods (at least 45 s) can effectively minimize fatigue effects and cumulative disturbances in the blood oxygenation response (Herold et al., 2018). When assessing brain network function, acquisitions longer than 4 min ensure stability and reliability of core metrics (Xu et al., 2023). The motor task requires 5–8 repetitions to obtain a stable signal, which may result in an excessively long total experimental time. The total experiment time is usually 15–20 min, with additional resting sessions when necessary to balance signal quality and patient attention duration. Most studies have adopted a “20-s task-30-s rest” pattern, repeated five times (Shen et al., 2024), to balance the total task time and data reliability. The design of experimental paradigms should account for individual differences in patients, including cognitive function, age, and pathological characteristics, to improve the scientific validity of fNIRS studies through individualized paradigm adjustments.
In event-related design, each trial is treated as a discrete event. The inter-stimulus interval (ISI) is typically set between 2 and 6 s (Geissler et al., 2021). It is recommended to use jittered interstimulus intervals or longer ISIs to ensure that the hemodynamic response function (HRF) for each event is not contaminated by adjacent stimuli. Randomized ISIs help reduce the effects of predictability and fatigue, thereby increasing statistical power (Jeong et al., 2025). The duration of the stimulus needs to match the characteristics of the stimulus conditions and the objectives of the experiment. For example, the Stroop task has a stimulus duration of 500 milliseconds (Li B. et al., 2024) to 2 s (Schroeter et al., 2002), while a more complex semantic task requires 5 s (Gilmore et al., 2021). This is because the former assesses quick reaction capabilities, while the latter requires participants to engage in semantic processing and information integration. The repetition frequency within a single trial is typically conducted 20 times (Geissler et al., 2021; Li B. et al., 2024) for each condition to ensure reliable signal averaging. The number of trials is determined by the specific research requirements and the sensitivity of the fNIRS system. Compared to block design, the event-related design of fNIRS paradigms is more suitable for exploring the instantaneous responses of brain regions to specific tasks. Therefore, event-related design is more appropriate for the developmental research of intervention technologies such as brain-computer interfaces and neurofeedback, which require real-time capabilities.
3.2.2 Stimulus presentation method
Stimulus modality selection should incorporate the perceptual characteristics of patients with stroke. Visual stimuli and auditory instructions are most commonly used in fNIRS studies. Auditory instructions should be short and clear to avoid increasing cognitive load (Potts et al., 2024; Peng et al., 2023). In practice, researchers tend to use multimodal stimulation (combining audio and visual) to improve task comprehension. Instruction design is a key aspect of fNIRS research in stroke. Patients with stroke often exhibit slowed information processing (Lugtmeijer et al., 2021); therefore, instruction design should follow the principle of simplification and use single-step instructions in the acute phase instead of multi-step instructions. Using the “demonstrate-practice-execute” model before starting a trial helps patients understand task requirements and significantly reduces the rate of data abandonment (Seitz et al., 2024). Feedback mechanisms are crucial to maintaining patient engagement (Shin and Chung, 2022; Kim et al., 2025), with immediate feedback facilitating correct task performance and delayed feedback favoring long-term learning (Palidis et al., 2025). Combining multiple feedback modalities can effectively stimulate the sensory system and promote the reconstruction of motor function and neuroplasticity (Rendos et al., 2021; Yuan et al., 2021; Noh et al., 2019; Matarasso et al., 2021). Combining neurofeedback tasks improves SMC activation in the affected hemisphere (Mihara et al., 2021), which is particularly important for rehabilitation in the chronic phase. Stimulus intensity can be designed in three ways: fixed difficulty, progressive enhancement, or adaptive design (Maier et al., 2019). Fixed difficulty is suitable for standardized assessment, progressive designs (e.g., from single to multiple joint movements) are suitable for training, and adaptive designs (adjusting difficulty based on real-time performance) are best for individualized rehabilitation (Ma et al., 2023; Matarasso et al., 2021; Rieke et al., 2020). Difficulty designs that generally aim for a 70–80% success rate strike a balance between challenge and frustration to optimize rehabilitation. fNIRS assessments can dynamically adjust task difficulty for blood oxygenation signals to avoid over-activation or under-response (Huo et al., 2022; Kohl et al., 2020).
3.2.3 Adaptive task design for stroke victims
Stroke, as a complex neurological disease, poses unique challenges for fNIRS task paradigm design. Patient motor dysfunction, cognitive impairment, and susceptibility to fatigue affect assessment feasibility, reduce data quality degradation, and trigger bias (Skau et al., 2021; Pan et al., 2019; Csípo et al., 2021; Takahashi et al., 2021). To address these challenges, adaptive task design can better accommodate patient functional differences while maintaining neuroscience rigor. The core concepts include: first, “feasibility first” to ensure that most patients can complete the assessment and avoid sample bias due to screening for functional level; second, “information maximization” to obtain as much valid neurological data as possible, even if the task is simplified. Achieving these two goals requires researchers to make strategic trade-offs in trial design. Three main adaptation strategies have been developed to address motor function limitations (Guo et al., 2022; Qu et al., 2025; Almulla et al., 2022; Su et al., 2023; Kotegawa and Teramoto, 2022): the healthy-side substitution task: by observing the effect of healthy-side movement on the affected neural network, the complex relationship between neural inhibition and facilitation across hemispheres is explored in depth, which helps to understand the potential mechanisms of neural reorganization after brain injury; the passive-motor paradigm: the activation of sensorimotor networks using an external assistive device, which provides an opportunity for severely limited motor ability; motor imagery tasks: indirectly assessing motor function and neuroplasticity through internal neural activation of motor preparation and planning networks. These three strategies cannot replace traditional tasks, but can reflect the key features of motor network reorganization from different perspectives.
Adaptive design for cognitive impairment is based on reducing non-target cognitive load (Bijarsari, 2021). By dynamically adjusting task difficulty so that the task is always in the patient's “comfort zone,” the frustration caused by a task that is too difficult and the ceiling effect caused by a task that is too easy are avoided. Extended reaction time and multimodal cues compensated for the decline in cognitive processing efficiency in both the temporal and perceptual dimensions, ensuring that the task was focused on cognitive functioning instead of processing speed or attention maintenance (Bello et al., 2021).
Regarding fatigue management, the adaptive design shifts from a “single continuous measure” to a “chunked cumulative assessment.” Short time chunking reduces the physical burden on the patient and improves signal quality by reducing interference from head movements and physiologic drift. Real-time load monitoring introduces the concept of adaptive measurement and dynamically adjusts task load based on neural indicators, which is superior to subjective reports (Matarasso et al., 2021; Kohl et al., 2020; Asgher et al., 2021). Gamification tasks combat fatigue by stimulating interest, enhancing ecological validity (Yu et al., 2022; Bae and Park, 2023). Individualized calibration and ability matching further safeguard the scientific and ethical nature of the study. Adaptive task design should focus on adaptive algorithm development, standard task library construction, and multimodal integration in the future to promote accurate personalized assessment.
3.3 Classification of task paradigms in stroke rehabilitation
3.3.1 Paradigms related to motor function rehabilitation
3.3.1.1 Active movement tasks
Active motor tasks, the most widely used paradigm in fNIRS stroke research (Table 2), are commonly used to explore the neural mechanisms of functional recovery of the upper limb (Xu G. et al., 2022; Lu et al., 2023; Ni et al., 2023; Xu R. et al., 2022; Du et al., 2022; Bonnal et al., 2024; Li C. et al., 2022). Most studies have assessed the activation status of the motor cortex on the affected side using a block design with specific motor paradigms (Lim and Eng, 2019; Yang et al., 2024; He et al., 2023; Mihara et al., 2021; Bonnal et al., 2023; Borrell et al., 2023; Zhao et al., 2024; Lim et al., 2021): the upper limb paradigm (grasping, finger-pairing, and picking up) may reflect fine-motor abilities with sensorimotor integration; the lower limb paradigm (standing, walking, and ankle dorsiflexion) focuses on balance and coordination, with the walking task having high ecological validity. Overall, these designs are simple and easy to standardize, but their relevance to everyday functioning remains to be strengthened.
Table 2. Classification of active movement task designs.
With a deeper understanding of neural plasticity, active motor task designs have evolved from single movements to task sequences, expanding from unilateral assessments to bilateral comparisons, and gradually introducing feedback interactions (Muller et al., 2024; Xu G. et al., 2022; Parker et al., 2024; Bae and Park, 2023; Du et al., 2022). VR/augmented reality (AR) technology enhances executive motivation through immediate visual feedback and, more importantly, creates closed-loop neurofeedback that integrates sensory, cognitive, and motor network assessment, breaking the limitations of traditional paradigms (Liu P. et al., 2024; Cui et al., 2025; Taguchi et al., 2022).
The design of active motor tasks should account for the patient's functional status and stage of recovery (Ma et al., 2023). Simple, low-intensity movements are preferred in the acute phase, and complexity can be gradually increased in the recovery phase; goal-oriented tasks enhance cortical activation (Lacerenza et al., 2023). This hierarchical design enhances data reliability and allows for continuous monitoring of the entire rehabilitation process. Moreover, safety is a primary consideration for patients with stroke performing motor tasks (especially walking paradigms), and suspension systems reduce the risk of falls (He et al., 2023). The active motor task paradigm should focus on individual differences, extend laboratory-standardized movements to functional tasks of daily living, and transition from single-assessment to longitudinal dynamic monitoring to enhance the clinical applicability of the fNIRS assessment and provide a strong neuroscientific support for precision rehabilitation.
3.3.1.2 Passive movement tasks
A passive motor task guided limb movements in patients with severe stroke by external forces and recorded cortical activation of sensory inputs, demonstrating that sensory pathways drive motor networks even in a paralyzed state (Cheng et al., 2024). This finding provides an important theoretical basis for early rehabilitation from a neuroscientific perspective-the integrity of sensory pathways may be a prerequisite for motor function recovery. Unlike active movements that primarily activate M1 and SMA, passive movements focus on the sensory cortex and premotor area (PMA) (Li et al., 2024). Robot-assisted technology has improved the standardization (Liu P. et al., 2024; Dai et al., 2024; Jiang et al., 2022) and data reliability (Xie et al., 2022b; Bonanno et al., 2023) of passive tasks, facilitating longitudinal monitoring of the rehabilitation process.
In clinical practice, passive motor tasks have a “dual role.” They assess the potential for sensory-motor integration and recovery and activate residual neural networks to prevent disuse atrophy through sustained passive activity. Future studies should focus on the neural mechanisms of passive adaptation to active remodeling to provide precise neurobiological markers for clinical rehabilitation grading. Passive motor tasks should be used as an alternative to active movement and guide patients from passive perception to active movement through sensory feedback to achieve true functional rehabilitation.
3.3.1.3 Motor imagery task
The unique value of the motor imagery task is that it builds a neural bridge of “intention-action” for patients who cannot perform actual movement but have preserved cognitive function (Hurst and Boe, 2022; Villa-Berges et al., 2023; Wang H. et al., 2023; Moran and O'Shea, 2020; Mehler et al., 2020). Unlike active or passive movements, MI does not depend on executive ability to activate the motor preparation and planning loop, which induces SMA/PMC and, in some patients, activation of the M1, the level of which strongly correlates with recovery potential (Mihara et al., 2021; Wang et al., 2025a).
Improvement in MI task effectiveness relies on fine-tuning the design. Studies have shown that externally-supported imagery (e.g., VR) is more effective than imagery alone (Choy et al., 2023; Kim D. H. et al., 2022); multisensory guidance (e.g., movement observation, sound rhythms) produces a stronger cortical response than a single instruction (Almulla et al., 2022; Errante et al., 2022; Eaves et al., 2024; Choi et al., 2022); graded-difficulty designs can be adapted to the imagery abilities of different patients (Ji et al., 2021; Wriessnegger et al., 2017); dominant hand tasks are more likely to induce strong activation, non-dominant hand imagery should be more vivid (Wang et al., 2025a). Regarding clinical translation, integrating MI tasks with brain-computer interface (BCI) systems is a new avenue for stroke rehabilitation (Batula et al., 2017; Wang et al., 2023; Ma Z. Z. et al., 2024; Khan et al., 2020). The fNIRS-based MI-BCI system helps patients adjust their imagery strategies through real-time feedback, reinforcing the activation of specific brain regions and activating more neural circuits than conventional training, especially the key pathways connecting intention and action (Lin et al., 2022; Liu X. et al., 2023).
Despite their advantages, MI tasks face challenges in objectively monitoring the quality of execution. Whether or not the patient actually performs the MI and the quality of imagery both directly affect the reliability of results. Studies have shown that false-positive feedback significantly reduces cortical activation during training in subjects, especially in contralateral motor areas, which can negatively impact cortical plasticity (Jeong et al., 2025). Therefore, a more accurate physiological index assessment system should be constructed using electromyography and other monitoring tools (Qin et al., 2025) to achieve quality control and enhance cortical activation and neuroplasticity. The MI task should be optimized through personalized customization, multimodal fusion, and closed-loop feedback to improve network activation and be linked with external assistive devices to build an “intention-execution-feedback” closed-loop, which provides dynamic support for stroke rehabilitation.
3.3.2 Cognitive tasks
Cognitive dysfunction is a common and prognostic problem after stroke, involving several core components such as working memory, inhibitory control, cognitive flexibility, and planning ability (Schumacher et al., 2019). fNIRS commonly uses three types of classical cognitive paradigms (Lu et al., 2025; Cheng X. P. et al., 2024; Udina et al., 2020): the verbal fluency task test (VFT) is used to assess verbal initiation and executive search ability; the n-back task targeting working memory capacity and updating function; the Stroop task assessing cognitive inhibitory control. Reduced PFC activation, interhemispheric balance dysregulation, and decreased efficiency of network integration are three neurophysiological features of executive dysfunction after stroke (Zou et al., 2023; Udina et al., 2020; Li X. et al., 2022). Results of cross-sectional studies have shown that the magnitude of PFC blood oxygenation signaling changes in cognitive tasks is significantly lower in patients with stroke than in healthy controls, and the attenuation is more pronounced with increased task difficulty (Kong et al., 2023; Csípo et al., 2021; Sunwoo et al., 2023; Sun et al., 2022; Huang et al., 2024). This underactivated performance was significantly correlated with clinical executive function scores (Ai et al., 2025; Li X. et al., 2023), supporting the feasibility of fNIRS as a marker of neurological function. Longitudinal studies have further found that patients in the early stages of stroke tend to show a compensatory response of PFC overactivation; as recovery progresses, the activation pattern gradually normalizes, and improvements in neurophysiological markers usually precede the recovery of behavioral function (Zou et al., 2023; Kong et al., 2023). This trajectory of change suggests a critical time window for rehabilitation interventions. To improve the ecological validity of measures and sensitivity to mild cognitive impairment, assessment paradigms are shifting from traditional single-tasks to more realistic designs such as multitask switching, complex problem solving, and simulation of daily activities in virtual reality environments to identify cognitive deficits at an earlier stage and guide rehabilitation interventions.
3.3.3 Language tasks
Language processing consists of four levels: phonological, syntactic, semantic, and pragmatic. The complexity of language dysfunction arises precisely from the multicomponent nature of language processing (Varkanitsa and Kiran, 2022; Stefaniak et al., 2021). fNIRS has a unique advantage in assessing post-stroke language function, as it allows patients to complete language tasks in a natural communicative environment, reducing the interference of motion and noise with language measurements that occurs with traditional neuroimaging techniques. Early studies validated the reliability of monitoring activation in classical language areas (Broca's and Wernicke's areas) through naming and word generation tasks (Gan et al., 2024; Hara et al., 2017). Subsequently, studies have progressively refined the assessment of the components of language functioning, developing specialized paradigms for different segments: naming tasks reflect the neural basis of patients' lexical extraction, semantic processing, and linguistic expression (Gilmore et al., 2021; Chang W. K. et al., 2022); semantic and phonological fluency tasks assess lexical extraction ability (Gilmore et al., 2021; Guo et al., 2025; Zhang et al., 2023; Fujii et al., 2021); bilingual switching tasks probe language control mechanisms (Farrukh et al., 2025). Recent studies have further revealed the critical role of the dorsal motor cortex in implicit speech, expanding the understanding of the functional connectivity of brain regions involved in language control (Si et al., 2021). In the clinical setting, patterns of PFC activation during acute-phase language tasks predict long-term recovery. Neuroindicators combined with traditional language assessment help create more accurate prognostic prediction models for early rehabilitation planning (Butler et al., 2020).
Current fNIRS language function assessment tasks still have significant limitations. Most paradigms are based on English, posing challenges in generalizing them to language systems, such as Chinese, that differ significantly in morphology, phonology, and semantic structure. Moreover, these tasks focus on the lexical and sentence levels and lack a holistic assessment at the dialog and chapter levels. The analysis focuses on activating local brain regions, neglecting the synergistic effect of the whole brain functional network during language processing. Future assessments of language function need improvements in several directions. For example, developing natural language tasks with higher ecological validity to cover multi-level processing contexts from words and sentences to real conversations and narratives; designing tasks specific to different language characteristics to better reflect the structure and function of each type of language; combining functional connectivity analyses to expand the scope of assessment from local brain area activation to the dynamics of the whole brain network. These improvements will enable the fNIRS language function assessment to integrate basic and clinical research and facilitate the development of personalized language rehabilitation programs based on neural mechanisms.
3.3.4 Swallowing tasks
Swallowing dysfunction is common after stroke and seriously affects the nutritional intake and quality of life in patients (Labeit et al., 2023). fNIRS is more commonly used for limb dysfunction and less for swallowing function (Gallois et al., 2022), probably because swallowing involves multiple regions of the brainstem and cortex and complex sensory-motor coordination, and the brainstem is a key control center. fNIRS is limited by its ability to detect deep brain regions (e.g., brainstem), posing challenges in capturing the activity of these core brain regions. Current fNIRS swallowing task paradigms typically include various designs, such as salivary, autonomous, and commanded swallowing (Ma X. et al., 2024; Wen et al., 2023; Wang et al., 2024; Matsuo et al., 2021) that reveal the functional state of swallowing-related neural networks by evoking specific cortical activation patterns. Recent studies have used a paradigm closer to everyday life, in which subjects are asked to complete sequential movements of gripping, chewing, and swallowing (Matsuo et al., 2021).
Studies have shown that swallowing tasks primarily activate the primary somatosensory cortex, motor cortex, frontal regions, and brainstem association areas (Ma X. et al., 2024; Wen et al., 2023). Swallowing task design should consider safety and feasibility, and rationally set the number of swallows, time interval, and command mode to avoid the risk of aspiration and fatigue. Moreover, similar cortical activation patterns are observed for imagined swallowing in patients who cannot perform the swallowing maneuver (Matsuo et al., 2021).
3.3.5 Dual-task design
The dual-task paradigm provides a measure closer to real-life scenarios for functional assessment of patients with stroke by simultaneously observing the interaction of cognitive and motor tasks (Stephens et al., 2023). Compared with a single task (Table 3), the dual-task paradigm can more sensitively detect subtle changes in early functional deficits and rehabilitation, revealing potential problems that are difficult to detect with single-task assessment (Chiaramonte et al., 2022; Lindberg et al., 2024; Ohzuno and Usuda, 2019). Based on the resource competition theory, cognitive and motor control systems share limited attentional resources, and the efficiency of their resource allocation determines dual-task performance (Strobach, 2020; Tsang et al., 2022; Strobach, 2024). Patients with stroke often show significant functional deficits and increased energy cost of walking (Cw) in dual-task conditions due to reduced processing resources and decreased allocation efficiency as a result of brain damage (Compagnat et al., 2023; Muci et al., 2022; Nonnekes et al., 2020).
Table 3. Summary of fNIRS single-task paradigms in stroke rehabilitation.
PFC was over-activated in dual-task conditions in patients with stroke and was significantly associated with execution costs (e.g., longer completion time, decreased accuracy), reflecting neural compensatory mechanisms. While different types of cognitive tasks affected motor performance differently, the interference produced by executive function tasks (e.g., working memory) was more pronounced, revealing the critical role of executive control networks in cognitive-motor integration. Dual-task training reduces PFC activation and improves behavioral performance, supporting the “neuroefficiency” hypothesis that training enhances the efficiency of resource allocation (Nosaka et al., 2023; Wang Q. et al., 2023; Sun et al., 2022; Compagnat et al., 2023; Bishnoi et al., 2021; Ding et al., 2024; Ou et al., 2024; Lim et al., 2022a).
Although the dual-task paradigm provides a more relevant measure for stroke rehabilitation assessment, it still suffers from the lack of standardization, limited comparability of results, unclear effects of heterogeneity in brain injury type on task performance, and neglect of whole-brain network synergies, as most studies concentrate on PFC. To this end, future work should focus on advancing the following directions: establishing a standardized dual-task assessment system with graded difficulty to meet the needs of patients with different levels of functioning; expanding the coverage of brain regions to achieve systematic research on the synergistic mechanism of multiple brain regions, such as motor area and parietal lobe; constructing a prediction model of dual-task performance and daily ability by combining with real-life functional indexes to enhance the value of clinical translation. The dual-task paradigm is a bridge that connects experimental research and clinical application and is expected to promote functional recovery in patients with stroke in complex environments. Future research should strengthen its ecological validity and practicality, helping the dual-task paradigm become an important basis for decision-making in stroke rehabilitation.
3.4 fNIRS and clinical metrics
fNIRS quantifies cortical activity by measuring the concentration changes of HbO and HbR, with HbO serving as the primary proxy for neurovascular coupling. Task-related hemodynamic responses are typically expressed as beta coefficients derived from the general linear model (GLM), which convolves the experimental design with an assumed hemodynamic response function (Yu H. et al., 2024; Chen et al., 2024). Spatially, ROIs such as the M1, SMA, or PFC are defined from channel arrays, and activation maps are generated using t-statistics or z-scores (Kim et al., 2023; Kim H. et al., 2022). Functional connectivity (FC) is computed via Pearson correlations (Li H. et al., 2023; Kong et al., 2023) or wavelet phase coherence (Liu L. et al., 2022) between pairs of channels or ROIs, while effective connectivity (EC) can be inferred with Granger causality (Bu et al., 2023; Zou et al., 2024) or transfer entropy (Jian et al., 2021b) to assess directional information flow. Finally, lateralization indices (LI) quantify hemispheric asymmetry (Chen Y. F. et al., 2023; Yuan et al., 2022), and graph-theory metrics (e.g., global efficiency, clustering coefficient) summarize large-scale network properties across the fNIRS-derived connectome (Huo et al., 2023; Yuan et al., 2022; Xu G. et al., 2022).
Clinical metrics in the literature systematically quantify motor, cognitive, and daily-life outcomes, anchored by the Fugl-Meyer Assessment (FMA) for limb motor recovery (Chatterjee et al., 2019; Jian et al., 2021a; Tamashiro et al., 2019), the NIH Stroke Scale (NIHSS) for neurologic deficit (Huo et al., 2023; Yang et al., 2024), and the Barthel Index for activities of daily living (Liu L. et al., 2022). Complementary scales include the Montreal Cognitive Assessment (MoCA) for cognition (Liu Y. et al., 2024), the Western Aphasia Battery (WAB) for language (Xie et al., 2022a; Gilmore et al., 2021), and specialized batteries such as the 10-Meter Walk Test for gait (Chatterjee et al., 2019; Lee A. et al., 2020) and the penetration-aspiration scale (PAS) for swallowing safety (Ma X. et al., 2024; Wen et al., 2023). These validated instruments provide objective benchmarks that enable clinicians and researchers to track progress, stratify patients, and correlate behavioral gains with underlying neuroplasticity measured by fNIRS.
4 Future research directions
Current fNIRS paradigms often rely on overly simplified tasks that lack ecological validity, limiting real-world relevance. Most fNIRS paradigms still employ laboratory-centric tasks—such as single-joint finger taps or word lists—that poorly reflect the multisensory, goal-directed demands of daily life after stroke. This ecological shortfall is compounded by low sensitivity: fixed block lengths and uniform difficulty can mask subtle, clinically relevant changes in patients with heterogeneous lesion patterns. Robustness is weakened by uncontrolled motion artifacts and by the absence of harmonized optode montages, leading to high inter-site variability. Furthermore, language paradigms are almost exclusively English-centric, limiting cross-linguistic validity. Finally, few protocols incorporate real-time physiological noise suppression, so cardiac or respiratory drift can masquerade as neural signal. Until these validity, sensitivity, and robustness issues are explicitly mitigated, clinical uptake will remain tentative.
Future fNIRS research on stroke rehabilitation task paradigms should focus on breaking through the following directions: first, to address the highly heterogeneous nature of functional impairments in stroke, it is necessary to develop a more intelligent and adaptive task paradigm design and realize dynamic and personalized task regulation based on individual neurological functional status using artificial intelligence and big data technology. Second, due to its limited spatial resolution (~2–3 cm) and shallow penetration depth, fNIRS cannot resolve activity in small or deep cortical regions, its millisecond-level sluggishness makes it less suitable than electroencephalogram (EEG) for tracking rapid neural dynamics. A multidimensional and cross-modal neurological function assessment system should be constructed, integrating fNIRS with EEG, electromyography, and other multi-source neuroimaging data to comprehensively analyze the brain network remodeling mechanism. Third, introducing VR and AR technologies and designing interactive tasks closer to daily life situations can improve the ecological validity of the task paradigm and enhance patient participation and training transfer effects. Fourth, promote the construction of a standardized task paradigm library, formulate unified task parameters, presentation standards, and data processing norms, and promote the comparability and result promotion of multicenter studies. Fifth, explore the innovative application of closed-loop neurofeedback and brain-computer interface technologies in rehabilitation tasks to achieve real-time neurofunctional monitoring and personalized feedback regulation. Finally, moving fNIRS from bench to bedside is slowed by its high cost, bulky cap/optode assemblies, and the need to train clinicians not only in neuroimaging theory but also in the nuanced interpretation of hemodynamic traces within noisy clinical environments. Beyond the clinic, its utility is further circumscribed by motion artifacts in ambulatory patients, limited depth penetration in adults with thick scalp/skull, and the absence of standardized normative databases for rapid bedside decision-making. Thus, multidisciplinary collaboration is needed to promote the translation of fNIRS technology from scientific research to clinical practice, which ultimately serves to improve functional recovery and quality of life in patients with stroke.
5 Conclusion
This paper systematically reviewed the latest research progress of fNIRS in stroke rehabilitation task paradigms, focusing on the design principles of motor, cognitive, language, swallowing, and dual-task paradigms and their neurological mechanism analysis value. This study provides a valuable reference for deepening the understanding of the fNIRS task paradigm in stroke rehabilitation neuroscience research and the subsequent design of more scientific, standardized, and clinically usable fNIRS assessment protocols. In the future, it is necessary to integrate multimodal technologies (e.g., real-time neurofeedback, VR) to construct a dynamic assessment framework and design a personalized paradigm based on injury pattern and recovery stage to optimize closed-loop of assessment-intervention-validation and promote an in-depth translation of the fNIRS from mechanism research to the clinical practice of precision rehabilitation.
Data availability statement
The datasets analyzed during the present study are publicly available in the published articles included in this systematic review. All references and data sources are listed in the manuscript and can be accessed through public databases such as PubMed and Embase.
Author contributions
YH: Conceptualization, Writing – original draft, Writing – review & editing. XZ: Conceptualization, Writing – original draft, Writing – review & editing. HZ: Data curation, Writing – review & editing. SL: Investigation, Writing – review & editing. JS: Data curation, Writing – review & editing. ZC: Data curation, Writing – review & editing. QF: Writing – review & editing. BL: Writing – review & editing. YS: Supervision, Writing – review & editing. FL: Supervision, Writing – review & editing. ZS: Funding acquisition, Writing – review & editing.
Funding
The author(s) declared that financial support was received for this work and/or its publication. This work was sponsored by Joint Program on Health Science and Technology Innovation of Hainan Province (WSJK2025QN049), Hainan Provincial Natural Science Foundation of China (No. 822QN497), and Project Supported by Hainan Province Clinical Medical Center (No. 0202067).
Acknowledgments
We thank Ratna B. at Wordvice Editing Service for Language Refinement that significantly improved grammatical accuracy and textual fluency in this work.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher's note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
Ai, Y., Zhang, Y., Zheng, F., Hu, H., Yin, M., Ye, Z., et al. (2025). Important role of the right hemisphere in post-stroke cognitive impairment: a functional near-infrared spectroscopy study. Neurophotonics 12:15008. doi: 10.1117/1.NPh.12.1.015008
Almulla, L., Al-Naib, I., Ateeq, I. S., and Althobaiti, M. (2022). Observation and motor imagery balance tasks evaluation: an fNIRS feasibility study. PLoS ONE 17:e0265898. doi: 10.1371/journal.pone.0265898
Asgher, U., Khan, M. J., Asif Nizami, M. H., Khalil, K., Ahmad, R., Ayaz, Y., et al. (2021). Motor training using mental workload (MWL) with an assistive soft exoskeleton system: a functional near-infrared spectroscopy (fNIRS) study for brain–machine interface (BMI). Front. Neurorobotics 15:605751. doi: 10.3389/fnbot.2021.605751
Bae, S., and Park, H. S. (2023). Development of immersive virtual reality-based hand rehabilitation system using a gesture-controlled rhythm game with vibrotactile feedback: an fNIRS pilot study. IEEE Trans. Neural Syst. Rehabil. Eng. 31, 3732–3743. doi: 10.1109/TNSRE.2023.3312336
Bai, Z., and Fong, K. N. K. (2020). “remind-to-move” treatment enhanced activation of the primary motor cortex in patients with stroke. Brain Topogr. 33, 275–283. doi: 10.1007/s10548-020-00756-7
Batula, A. M., Mark, J. A., Kim, Y. E., and Ayaz, H. (2017). Comparison of brain activation during motor imagery and motor movement using fNIRS. Comput. Intell. Neurosci. 2017:5491296. doi: 10.1155/2017/5491296
Bello, U. M., Chan, C. C. H., and Winser, S. J. (2021). Task complexity and image clarity facilitate motor and visuo-motor activities in mirror therapy in post-stroke patients. Front. Neurol. 12:722846. doi: 10.3389/fneur.2021.722846
Bello, U. M., Winser, S. J., and Chan, C. C. H. (2020). Does task complexity influence motor facilitation and visuo-motor memory during mirror therapy in post-stroke patients? Med. Hypotheses 138:109590. doi: 10.1016/j.mehy.2020.109590
Bijarsari, E. (2021). A current view on dual-task paradigms and their limitations to capture cognitive load. Front. Psychol. 12:648586. doi: 10.3389/fpsyg.2021.648586
Bishnoi, A., Holtzer, R., and Hernandez, M. E. (2021). Brain activation changes while walking in adults with and without neurological disease: systematic review and meta-analysis of functional near-infrared spectroscopy studies. Brain Sci. 11, 1–22. doi: 10.3390/brainsci11030291
Bonanno, L., Cannuli, A., Pignolo, L., Marino, S., Quartarone, A., Calabrò, R. S., et al. (2023). Neural plasticity changes induced by motor robotic rehabilitation in stroke patients: the contribution of functional neuroimaging. Bioengineering 10:990. doi: 10.3390/bioengineering10080990
Bonnal, J., Ozsancak, C., Monnet, F., Valery, A., Prieur, F., and Auzou, P. (2023). Neural substrates for hand and shoulder movement in healthy adults: a functional near infrared spectroscopy study. Brain Topogr. 36, 447–458. doi: 10.1007/s10548-023-00972-x
Bonnal, J., Ozsancak, C., Prieur, F., and Auzou, P. (2024). Video mirror feedback induces more extensive brain activation compared to the mirror box: an fNIRS study in healthy adults. J. Neuroeng. Rehabil. 21:78. doi: 10.1186/s12984-024-01374-1
Borrell, J. A., Fraser, K., Manattu, A. K., and Zuniga, J. M. (2023). Laterality index calculations in a control study of functional near infrared spectroscopy. Brain Topogr. 36, 210–222. doi: 10.1007/s10548-023-00942-3
Bu, L., Qu, J., Zhao, L., Zhang, Y., and Wang, Y. (2023). A neuroergonomic approach to assessing motor performance in stroke patients using fNIRS and behavioral data. Appl. Ergon. 109:103979. doi: 10.1016/j.apergo.2023.103979
Butler, L. K., Kiran, S., and Tager-Flusberg, H. (2020). Functional near-infrared spectroscopy in the study of speech and language impairment across the life span: a systematic review. Am. J. Speech-Lang. Pathol. 29, 1674–1701. doi: 10.1044/2020_AJSLP-19-00050
Caliandro, P., Molteni, F., Simbolotti, C., Guanziroli, E., Iacovelli, C., Reale, G., et al. (2020). Exoskeleton-assisted gait in chronic stroke: an EMG and functional near-infrared spectroscopy study of muscle activation patterns and prefrontal cortex activity. Clin. Neurophysiol. 131, 1775–1781. doi: 10.1016/j.clinph.2020.04.158
Cao, H., Chen, O. Y., McEwen, S. C., Forsyth, J. K., Gee, D. G., Bearden, C. E., et al. (2021). Cross-paradigm connectivity: reliability, stability, and utility. Brain Imaging Behav. 15, 614–629. doi: 10.1007/s11682-020-00272-z
Chang, P. W., Lu, C. F., Chang, S. T., and Tsai, P. Y. (2022). Functional near-infrared spectroscopy as a target navigator for rTMS modulation in patients with hemiplegia: a randomized control study. Neurol. Ther. 11, 103–121. doi: 10.1007/s40120-021-00300-0
Chang, W. K., Park, J., Lee, J. Y., Cho, S., Lee, J., Kim, W. S., et al. (2022). Functional network changes after high-frequency rTMS over the most activated speech-related area combined with speech therapy in chronic stroke with non-fluent aphasia. Front. Neurol. 13:690048. doi: 10.3389/fneur.2022.690048
Chatterjee, S. A., Fox, E. J., Daly, J. J., Rose, D. K., Wu, S. S., Christou, E. A., et al. (2019). Interpreting prefrontal recruitment during walking after stroke: influence of individual differences in mobility and cognitive function. Front. Hum. Neurosci. 13:194. doi: 10.3389/fnhum.2019.00194
Chen, N., Qiu, X., Hua, Y., Hu, J., and Bai, Y. (2023). Effects of sequential inhibitory and facilitatory repetitive transcranial magnetic stimulation on neurological and functional recovery of a patient with chronic stroke: a case report and literature review. Front. Neurol. 14:1064718. doi: 10.3389/fneur.2023.1064718
Chen, S., Mao, M., Zhu, G., Chen, Y., Qiu, Y., Ye, B., et al. (2024). Cortical activity in patients with high-functioning ischemic stroke during the purdue pegboard test: insights into bimanual coordinated fine motor skills with functional near-infrared spectroscopy. Neural Regener. Res. 19, 1098–1104. doi: 10.4103/1673-5374.385312
Chen, S., Zhang, X., Chen, X., Zhou, Z., Cong, W., Chong, K., et al. (2023). The assessment of interhemispheric imbalance using functional near-infrared spectroscopic and transcranial magnetic stimulation for predicting motor outcome after stroke. Front. Neurosci. 17:1231693. doi: 10.3389/fnins.2023.1231693
Chen, W. L., Wagner, J., Heugel, N., Sugar, J., Lee, Y. W., Conant, L., et al. (2020). Functional near-infrared spectroscopy and its clinical application in the field of neuroscience: advances and future directions. Front. Neurosci. 14:724. doi: 10.3389/fnins.2020.00724
Chen, Y. F., Mao, M. C., Zhu, G. Y., Sun, C. C., Zhao, J. W., He, H. X., et al. (2023). The changes of neuroactivity of tui na (Chinese massage) at hegu acupoint on sensorimotor cortex in stroke patients with upper limb motor dysfunction: a fNIRS study. BMC Complementary Med. Ther. 23:334. doi: 10.1186/s12906-023-04143-0
Cheng, C., Liu, T., Zhang, B., Wu, X., Song, Z., Zhao, Z., et al. (2024). Effects of robot-assisted hand function therapy on brain functional mechanisms: a synchronized study using fNIRS and sEMG. Front. Med. 11:1411616. doi: 10.3389/fmed.2024.1411616
Cheng, X. P., Wang, Z. D., Zhou, Y. Z., Zhan, L. Q., Wu, D., Xie, L. L., et al. (2024). Effect of tDCS combined with virtual reality for post-stroke cognitive impairment: a randomized controlled trial study protocol. BMC Complementary Med. Ther. 24:349. doi: 10.1186/s12906-024-04658-0
Chiaramonte, R., Bonfiglio, M., Leonforte, P., Coltraro, G. L., Guerrera, C. S., and Vecchio, M. (2022). Proprioceptive and dual-task training: the key of stroke rehabilitation, a systematic review. J. Funct. Morphol. Kinesiol. 7:53. doi: 10.3390/jfmk7030053
Choi, J. B., Yang, S. W., and Ma, S. R. (2022). The effect of action observation combined with motor imagery training on upper extremity function and corticospinal excitability in stroke patients: a randomized controlled trial. Int. J. Environ. Res. Public Health 19:12048. doi: 10.3390/ijerph191912048
Choy, C. S., Fang, Q., Neville, K., Ding, B., Kumar, A., Mahmoud, S. S., et al. (2023). Virtual reality and motor imagery for early post-stroke rehabilitation. Biomed. Eng. Online 22:66. doi: 10.1186/s12938-023-01124-9
Chu, M., Zhang, Y., Chen, J., Chen, W., Hong, Z., Zhang, Y., et al. (2022). Efficacy of intermittent theta-burst stimulation and transcranial direct current stimulation in treatment of post-stroke cognitive impairment. J. Integr. Neurosci. 21:130. doi: 10.31083/j.jin2105130
Chu, Q., Guo, X., Zhang, T., Huo, C., Zhang, X., Xu, G., et al. (2023). Stroke-related alterations in the brain's functional connectivity response associated with upper limb multi-joint linkage movement. Brain Sci. 13:338. doi: 10.3390/brainsci13020338
Collett, J., Fleming, M. K., Meester, D., Al-Yahya, E., Wade, D. T., Dennis, A., et al. (2021). Dual-task walking and automaticity after stroke: insights from a secondary analysis and imaging sub-study of a randomised controlled trial. Clin. Rehabil. 35, 1599–1610. doi: 10.1177/02692155211017360
Compagnat, M., Daviet, J. C., Hermand, E., Billot, M., Salle, J. Y., and Perrochon, A. (2023). Impact of a dual task on the energy cost of walking in individuals with subacute phase stroke. Brain Inj. 37, 114–121. doi: 10.1080/02699052.2023.2165153
Csípo, T., Tanoue, T., Lipécz, Á., Mukli, P., et al. (2021). Increased cognitive workload evokes greater neurovascular coupling responses in healthy young adults. PLoS ONE 16:e0250043. doi: 10.1371/journal.pone.0250043
Cui, Y., Cong, F., Zeng, M., and Wang, J. (2025). Effects and mechanisms of synchronous virtual reality action observation and electrical stimulation on upper extremity motor function and activities of daily living in patients with stroke: a protocol for a randomized controlled trial. Front. Neurol. 16:1499178. doi: 10.3389/fneur.2025.1499178
Dai, L., Zhang, W., Zhang, H., Fang, L., Chen, J., Li, X., et al. (2024). Effects of robot-assisted upper limb training combined with intermittent theta burst stimulation (iTBS) on cortical activation in stroke patients: a functional near-infrared spectroscopy study. NeuroRehabilitation 54, 421–434. doi: 10.3233/NRE-230355
Defenderfer, J., Kerr-German, A., Hedrick, M., and Buss, A. T. (2017). Investigating the role of temporal lobe activation in speech perception accuracy with normal hearing adults: an event-related fNIRS study. Neuropsychologia 106, 31–41. doi: 10.1016/j.neuropsychologia.2017.09.004
Delorme, M., Vergotte, G., Perrey, S., Froger, J., and Laffont, I. (2019). Time course of sensorimotor cortex reorganization during upper extremity task accompanying motor recovery early after stroke: an fNIRS study. Restor. Neurol. Neurosci. 37, 207–218. doi: 10.3233/RNN-180877
Ding, Q., Ou, Z., Yao, S., Wu, C., Chen, J., Shen, J., et al. (2024). Cortical activation and brain network efficiency during dual tasks: an fNIRS study. Neuroimage 289:120545. doi: 10.1016/j.neuroimage.2024.120545
Du, Q., Luo, J., Cheng, Q., Wang, Y., and Guo, S. (2022). Vibrotactile enhancement in hand rehabilitation has a reinforcing effect on sensorimotor brain activities. Front. Neurosci. 16:935827. doi: 10.3389/fnins.2022.935827
Eaves, D. L., Hodges, N. J., Buckingham, G., Buccino, G., and Vogt, S. (2024). Enhancing motor imagery practice using synchronous action observation. Psychol. Res. 88, 1891–1907. doi: 10.1007/s00426-022-01768-7
Errante, A., Saviola, D., Cantoni, M., Iannuzzelli, K., Ziccarelli, S., Togni, F., et al. (2022). Effectiveness of action observation therapy based on virtual reality technology in the motor rehabilitation of paretic stroke patients: a randomized clinical trial. BMC Neurol. 22:109. doi: 10.1186/s12883-022-02640-2
Farrukh, F., Nazeer, H., Minhas, H. S., Naseer, N., and Noori, F. M. (2025). Assessing multilingual speakers' language processing through functional near-infrared spectroscopy (fNIRS). Behav. Brain Res. 484:115485. doi: 10.1016/j.bbr.2025.115485
Fu, X., Li, H., Yang, W., Li, X., Lu, L., Guo, H., et al. (2023). Electroacupuncture at HT5 + GB20 promotes brain remodeling and significantly improves swallowing function in patients with stroke. Front. Neurosci. 17:1274419. doi: 10.3389/fnins.2023.1274419
Fujii, M., Tanigo, K., Yamamoto, H., Kikugawa, K., Shirakawa, M., Ohgushi, M., et al. (2021). A case of dysgraphia after cerebellar infarction where functional NIRS guided the task aimed at activating the hypoperfused region. Case Rep. Neurol. Med. 2021:6612541. doi: 10.1155/2021/6612541
Gallois, Y., Neveu, F., Gabas, M., Cormary, X., Gaillard, P., Verin, E., et al. (2022). Can swallowing cerebral neurophysiology be evaluated during ecological food intake conditions? a systematic literature review. J. Clin. Med. 11:5480. doi: 10.3390/jcm11185480
Gan, L., Huang, L., Zhang, Y., Yang, X., Li, L., Meng, L., et al. (2024). Effects of low-frequency rTMS combined with speech and language therapy on broca's aphasia in subacute stroke patients. Front. Neurol. 15:1473254. doi: 10.3389/fneur.2024.1473254
GBD 2019 Stroke (2021). Global, regional, and national burden of stroke and its risk factors, 1990–2019: a systematic analysis for the global burden of disease study 2019. Lancet Neurol. 20, 795–820. doi: 10.1016/S1474-4422(21)00252-0
Geissler, C. F., Frings, C., and Moeller, B. (2021). Illuminating the prefrontal neural correlates of action sequence disassembling in response–response binding. Sci. Rep. 11:22856. doi: 10.1038/s41598-021-02247-6
Gilmore, N., Yucel, M. A., Li, X., Boas, D. A., and Kiran, S. (2021). Investigating language and domain-general processing in neurotypicals and individuals with aphasia - a functional near-infrared spectroscopy pilot study. Front. Hum. Neurosci. 15:728151. doi: 10.3389/fnhum.2021.728151
Guo, C., Sui, Y., Xu, S., Zhuang, R., Zhang, M., Zhu, S., et al. (2022). Contralaterally controlled neuromuscular electrical stimulation-induced changes in functional connectivity in patients with stroke assessed using functional near-infrared spectroscopy. Front. Neural Circuits 16:955728. doi: 10.3389/fncir.2022.955728
Guo, Z., Xu, L., Tan, W., and Chen, F. (2025). Impact of generation rate of speech imagery on neural activity and BCI decoding performance: a fNIRS study. IEEE Trans. Neural Syst. Rehabil. Eng. 33, 1180–1190. doi: 10.1109/TNSRE.2025.3552606
Hara, T., Abo, M., Kakita, K., Mori, Y., Yoshida, M., and Sasaki, N. (2017). The effect of selective transcranial magnetic stimulation with functional near-infrared spectroscopy and intensive speech therapy on individuals with post-stroke aphasia. Eur. Neurol. 77, 186–194. doi: 10.1159/000457901
He, X., Lei, L., Yu, G., Lin, X., Sun, Q., and Chen, S. (2023). Asymmetric cortical activation in healthy and hemiplegic individuals during walking: a functional near-infrared spectroscopy neuroimaging study. Front. Neurol. 13:1044982. doi: 10.3389/fneur.2022.1044982
Heiberg, A. V., Simonsen, S. A., Schytz, H. W., and Iversen, H. K. (2023). Cortical hemodynamic response during cognitive stroop test in acute stroke patients assessed by fNIRS. NeuroRehabilitation 52, 199–217. doi: 10.3233/NRE-220171
Herold, F., Wiegel, P., Scholkmann, F., and Müller, N. (2018). Applications of functional near-infrared spectroscopy (fNIRS) neuroimaging in exercise–cognition science: a systematic, methodology-focused review. J. Clin. Med. 7:466. doi: 10.3390/jcm7120466
Herold, F., Wiegel, P., Scholkmann, F., Thiers, A., Hamacher, D., and Schega, L. (2017). Functional near-infrared spectroscopy in movement science: a systematic review on cortical activity in postural and walking tasks. Neurophotonics 4:041403. doi: 10.1117/1.NPh.4.4.041403
Huang, Y. H., Chen, W. Y., Liu, Y. H., Li, T. Y., Lin, C. P., Cheong, P. L., et al. (2024). Mild cognitive impairment estimation based on functional near-infrared spectroscopy. J. Biophotonics 17:e202300251. doi: 10.1002/jbio.202300251
Huo, C., Sun, Z., Xu, G., Li, X., Xie, H., Song, Y., et al. (2022). fNIRS-based brain functional response to robot-assisted training for upper-limb in stroke patients with hemiplegia. Front. Aging Neurosci. 14:1060734. doi: 10.3389/fnagi.2022.1060734
Huo, C., Xu, G., Sun, A., Xie, H., Hu, X., Li, W., et al. (2023). Cortical response induced by task-oriented training of the upper limb in subacute stroke patients as assessed by functional near-infrared spectroscopy. J. Biophotonics 16:e202200228. doi: 10.1002/jbio.202200228
Hurst, A. J., and Boe, S. G. (2022). Imagining the way forward: a review of contemporary motor imagery theory. Front. Hum. Neurosci. 16:1033493. doi: 10.3389/fnhum.2022.1033493
Jalalvandi, M., Sharini, H., Shafaghi, L., and Alam, N. R. (2024). Deciphering brain activation during wrist movements: comparative fMRI and fNIRS analysis of active, passive, and imagery states. Exp. Brain Res. 243:36. doi: 10.1007/s00221-024-06977-7
Jeong, H., Song, M., Jang, S. H., and Kim, J. (2025). Investigating the cortical effect of false positive feedback on motor learning in motor imagery based rehabilitative BCI training. J. Neuroeng. Rehabil. 22:61. doi: 10.1186/s12984-025-01597-w
Ji, E. K., Wang, H. H., Jung, S. J., Lee, K. B., Kim, J. S., Jo, L., et al. (2021). Graded motor imagery training as a home exercise program for upper limb motor function in patients with chronic stroke: a randomized controlled trial. Medicine 100:e24351. doi: 10.1097/MD.0000000000024351
Jian, C., Deng, L., Liu, H., Yan, T., Wang, X., and Song, R. (2021a). Modulating and restoring inter-muscular coordination in stroke patients using two-dimensional myoelectric computer interface: a cross-sectional and longitudinal study. J. Neural Eng. 18. doi: 10.1088/1741-2552/abc29a
Jian, C., Liu, H., Deng, L., Wang, X., Yan, T., and Song, R. (2021b). Stroke-induced alteration in multi-layer information transmission of cortico-motor system during elbow isometric contraction modulated by myoelectric-controlled interfaces. J. Neural Eng. 18. doi: 10.1088/1741-2552/ac18ae
Jiang, Y. C., Ma, R., Qi, S., Ge, S., Sun, Z., Li, Y., et al. (2022). Characterization of bimanual cyclical tasks from single-trial EEG-fNIRS measurements. IEEE Trans. Neural Syst. Rehabil. Eng. 30, 146–156. doi: 10.1109/TNSRE.2022.3144216
Khan, M. A., Das, R., Iversen, H. K., and Puthusserypady, S. (2020). Review on motor imagery based BCI systems for upper limb post-stroke neurorehabilitation: from designing to application. Comput. Biol. Med. 123:103843. doi: 10.1016/j.compbiomed.2020.103843
Kim, D. H., Lee, K. D., Bulea, T. C., and Park, H. S. (2022). Increasing motor cortex activation during grasping via novel robotic mirror hand therapy: a pilot fNIRS study. J. Neuroeng. Rehabil. 19:8. doi: 10.1186/s12984-022-00988-7
Kim, H., Kim, J., Lee, G., Lee, J., and Kim, Y. H. (2022). Task-related hemodynamic changes induced by high-definition transcranial direct current stimulation in chronic stroke patients: an uncontrolled pilot fNIRS study. Brain Sci. 12:453. doi: 10.3390/brainsci12040453
Kim, H., Lee, G., Lee, J., and Kim, Y. H. (2023). Alterations in learning-related cortical activation and functional connectivity by high-definition transcranial direct current stimulation after stroke: an fNIRS study. Front. Neurosci. 17:1189420. doi: 10.3389/fnins.2023.1189420
Kim, J., Kim, E., Lee, S. H., Lee, G., and Kim, Y. H. (2024). Use of cortical hemodynamic responses in digital therapeutics for upper limb rehabilitation in patients with stroke. J. Neuroeng. Rehabil. 21:115. doi: 10.1186/s12984-024-01404-y
Kim, M. S., Park, H., Kwon, I., An, K. O., Kim, H., Park, G., et al. (2025). Efficacy of brain-computer interface training with motor imagery-contingent feedback in improving upper limb function and neuroplasticity among persons with chronic stroke: a double-blinded, parallel-group, randomized controlled trial. J. Neuroeng. Rehabil. 22:1. doi: 10.1186/s12984-024-01535-2
Kinoshita, S., Tamashiro, H., Okamoto, T., Urushidani, N., and Abo, M. (2019). Association between imbalance of cortical brain activity and successful motor recovery in sub-acute stroke patients with upper limb hemiparesis: a functional near-infrared spectroscopy study. Neuroreport 30, 822–827. doi: 10.1097/WNR.0000000000001283
Kohl, S. H., Mehler, D. M. A., Lührs, M., Thibault, R. T., Konrad, K., and Sorger, B. (2020). The potential of functional near-infrared spectroscopy-based neurofeedback-a systematic review and recommendations for best practice. Front. Neurosci. 14:594. doi: 10.3389/fnins.2020.00594
Kong, Y., Peng, W., Li, J., Zhu, C., Zhang, C., and Fan, Y. (2023). Alteration in brain functional connectivity in patients with post-stroke cognitive impairment during memory task: a fNIRS study. J. Stroke Cerebrovasc. Dis. 32:107280. doi: 10.1016/j.jstrokecerebrovasdis.2023.107280
Kotegawa, K., and Teramoto, W. (2022). Association of executive function capacity with gait motor imagery ability and PFC activity: an fNIRS study. Neurosci. Lett. 766:136350. doi: 10.1016/j.neulet.2021.136350
Kotegawa, K., Yasumura, A., and Teramoto, W. (2020). Activity in the prefrontal cortex during motor imagery of precision gait: an fNIRS study. Exp. Brain Res. 238, 221–228. doi: 10.1007/s00221-019-05706-9
Labeit, B., Michou, E., Hamdy, S., Trapl-Grundschober, M., Suntrup-Krueger, S., Muhle, P., et al. (2023). The assessment of dysphagia after stroke: state of the art and future directions. Lancet Neurol. 22, 858–870. doi: 10.1016/S1474-4422(23)00153-9
Lacerenza, M., Frabasile, L., Buttafava, M., Spinelli, L., Bassani, E., Micheloni, F., et al. (2023). Motor cortex hemodynamic response to goal-oriented and non-goal-oriented tasks in healthy subjects. Front. Neurosci. 17:1202705. doi: 10.3389/fnins.2023.1202705
Lee, A., Kim, H., Kim, J., Choi, D. S., Jung, J. H., Lee, J., et al. (2020). Modulating effects of whole-body vibration on cortical activity and gait function in chronic stroke patients. Brain Neurorehabilit. 13:e12. doi: 10.12786/bn.2020.13.e12
Lee, S. H., Lee, H. J., Shim, Y., Chang, W. H., Choi, B. O., Ryu, G. H., et al. (2020). Wearable hip-assist robot modulates cortical activation during gait in stroke patients: a functional near-infrared spectroscopy study. J. Neuroeng. Rehabil. 17:145. doi: 10.1186/s12984-020-00777-0
Li, B., Li, M., Xia, J., Jin, H., Dong, S., and Luo, J. (2024). Hybrid integrated wearable patch for brain EEG-fNIRS monitoring. Sensors 24:4847. doi: 10.3390/s24154847
Li, C., Chen, Y., Tu, S., Lin, J., Lin, Y., Xu, S., et al. (2024). Dual-tDCS combined with sensorimotor training promotes upper limb function in subacute stroke patients: a randomized, double-blinded, sham-controlled study. CNS Neurosci. Ther. 30:e14530. doi: 10.1111/cns.14530
Li, C., Wong, Y., Langhammer, B., Huang, F., Du, X., Wang, Y., et al. (2022). A study of dynamic hand orthosis combined with unilateral task-oriented training in subacute stroke: a functional near-infrared spectroscopy case series. Front. Neurol. 13:907186. doi: 10.3389/fneur.2022.907186
Li, H., Fu, X., Lu, L., Guo, H., Yang, W., Guo, K., et al. (2023). Upper limb intelligent feedback robot training significantly activates the cerebral cortex and promotes the functional connectivity of the cerebral cortex in patients with stroke: a functional near-infrared spectroscopy study. Front. Neurol. 14:1042254. doi: 10.3389/fneur.2023.1042254
Li, W., Zhu, G., Jiang, Y., Miao, C., Zhang, G., and Xu, D. (2024). Cortical response characteristics of passive, active, and resistance movements: a multi-channel fNRIS study. Front. Hum. Neurosci. 18:1419140. doi: 10.3389/fnhum.2024.1419140
Li, X., Fang, F., Li, R., and Zhang, Y. (2022). Functional brain controllability alterations in stroke. Front. Bioeng. Biotechnol. 10:925970. doi: 10.3389/fbioe.2022.925970
Li, X., Huang, F., Guo, T., Feng, M., and Li, S. (2023). The continuous performance test aids the diagnosis of post-stroke cognitive impairment in patients with right hemisphere damage. Front. Neurol. 14:1173004. doi: 10.3389/fneur.2023.1173004
Liang, J., Song, Y., Belkacem, A. N., Li, F., Liu, S., Chen, X., et al. (2022). Prediction of balance function for stroke based on EEG and fNIRS features during ankle dorsiflexion. Front. Neurosci. 16:968928. doi: 10.3389/fnins.2022.968928
Lim, S. B., and Eng, J. J. (2019). Increased sensorimotor cortex activation with decreased motor performance during functional upper extremity tasks poststroke. J. Neurol. Phys. Ther. 43, 141. doi: 10.1097/NPT.0000000000000277
Lim, S. B., Louie, D. R., Peters, S., Liu-Ambrose, T., Boyd, L. A., and Eng, J. J. (2021). Brain activity during real-time walking and with walking interventions after stroke: a systematic review. J. Neuroeng. Rehabil. 18:8. doi: 10.1186/s12984-020-00797-w
Lim, S. B., Peters, S., Yang, C. L., Boyd, L. A., Liu-Ambrose, T., and Eng, J. J. (2022a). Frontal, sensorimotor, and posterior parietal regions are involved in dual-task walking after stroke. Front. Neurol. 13:904145. doi: 10.3389/fneur.2022.904145
Lim, S. B., Yang, C. L., Peters, S., Liu-Ambrose, T., Boyd, L. A., and Eng, J. J. (2022b). Phase-dependent brain activation of the frontal and parietal regions during walking after stroke - an fNIRS study. Front. Neurol. 13:904722. doi: 10.3389/fneur.2022.904722
Lin, B., Ni, J., Xiong, X., Zhang, L., Song, J., Wang, M., et al. (2025). Language function improvement and cortical activity alteration using scalp acupuncture coupled with speech-language training in post-stroke aphasia: a randomised controlled study. Complement. Ther. Med. 89:103137. doi: 10.1016/j.ctim.2025.103137
Lin, Q., Zhang, Y., Zhang, Y., Zhuang, W., Zhao, B., Ke, X., et al. (2022). The frequency effect of the motor imagery brain computer interface training on cortical response in healthy subjects: a randomized clinical trial of functional near-infrared spectroscopy study. Front. Neurosci. 16:810553. doi: 10.3389/fnins.2022.810553
Lindberg, P. G., AmirShemiraniha, N., Krewer, C., Maier, M. A., and Hermsdörfer, J. (2024). Increased dual-task interference during upper limb movements in stroke exceeding that found in aging - a systematic review and meta-analysis. Front. Neurol. 15:1375152. doi: 10.3389/fneur.2024.1375152
Liu, H., Peng, Y., Liu, Z., Wen, X., Li, F., Zhong, L., et al. (2022). Hemodynamic signal changes and swallowing improvement of repetitive transcranial magnetic stimulation on stroke patients with dysphagia: a randomized controlled study. Front. Neurol. 13:918974. doi: 10.3389/fneur.2022.918974
Liu, L., Jin, M., Zhang, L., Zhang, Q., Hu, D., Jin, L., et al. (2022). Brain-computer interface-robot training enhances upper extremity performance and changes the cortical activation in stroke patients: a functional near-infrared spectroscopy study. Front. Neurosci. 16:809657. doi: 10.3389/fnins.2022.809657
Liu, P., Yang, X., Han, F., Peng, G., Li, Q., Huang, L., et al. (2024). Brain activation pattern caused by soft rehabilitation glove and virtual reality scenes: a pilot fNIRS study. IEEE Trans. Neural Syst. Rehabil. Eng. 32, 3848–3857. doi: 10.1109/TNSRE.2024.3482470
Liu, X., Zhang, W., Li, W., Zhang, S., Lv, P., and Yin, Y. (2023). Effects of motor imagery based brain-computer interface on upper limb function and attention in stroke patients with hemiplegia: a randomized controlled trial. BMC Neurol. 23:136. doi: 10.1186/s12883-023-03150-5
Liu, Y., Luo, J., Fang, J., Yin, M., Cao, J., Zhang, S., et al. (2023). Screening diagnosis of executive dysfunction after ischemic stroke and the effects of transcranial magnetic stimulation: a prospective functional near-infrared spectroscopy study. CNS Neurosci. Ther. 29, 1561–1570. doi: 10.1111/cns.14118
Liu, Y., Zhong, Z., Chen, J., Kuo, H., Chen, X., Wang, P., et al. (2024). Brain activation patterns in patients with post-stroke cognitive impairment during working memory task: a functional near-infrared spectroscopy study. Front. Neurol. 15:1419128. doi: 10.3389/fneur.2024.1419128
Lu, C., Wang, M., Zhan, L., and Lu, M. (2025). Unveiling cognitive interference: FNIRS insights into poststroke aphasia during stroop tasks. Neural Plast. 2025:1456201. doi: 10.1155/np/1456201
Lu, W., Jin, X., Chen, J., Liu, G., Wang, P., Hu, X., et al. (2023). Prefrontal cortex activity of active motion, cyclic electrical muscle stimulation, assisted motion, and imagery of wrist extension in stroke using fNIRS. J. Stroke Cerebrovasc. Dis. 32:107456. doi: 10.1016/j.jstrokecerebrovasdis.2023.107456
Lugtmeijer, S., Lammers, N. A., de Haan, E. H. F., de Leeuw, F. E., and Kessels, R. P. C. (2021). Post-stroke working memory dysfunction: a meta-analysis and systematic review. Neuropsychol. Rev. 31, 202–219. doi: 10.1007/s11065-020-09462-4
Luke, R., Larson, E., Shader, M. J., Innes-Brown, H., Van Yper, L., Lee, A. K. C., et al. (2021). Analysis methods for measuring passive auditory fNIRS responses generated by a block-design paradigm. Neurophotonics 8:25008. doi: 10.1117/1.NPh.8.2.025008
Ma, H., Zhai, Y., Xu, Z., Fan, S., Wu, X., Xu, J., et al. (2022). Increased cerebral cortex activation in stroke patients during electrical stimulation of cerebellar fastigial nucleus with functional near-infrared spectroscopy. Front. Neurosci. 16:895237. doi: 10.3389/fnins.2022.895237
Ma, X., Peng, Y., Zhong, L., Li, F., Tang, Z., Bao, X., et al. (2024). Hemodynamic signal changes during volitional swallowing in dysphagia patients with different unilateral hemispheric stroke and brainstem stroke: a near-infrared spectroscopy study. Brain Res. Bull. 207:110880. doi: 10.1016/j.brainresbull.2024.110880
Ma, Y., Xie, D., Yu, Y., Yao, K., Zhang, S., Li, Q., et al. (2025). Differences in brain activation and connectivity during unaffected hand exercise in subacute and convalescent stroke patients. Neuroscience 565, 10–18. doi: 10.1016/j.neuroscience.2024.11.038
Ma, Y., Yu, Y., Gao, W., Hong, Y., and Shen, X. (2023). Cerebral hemodynamic changes during unaffected handgrip exercises in stroke patients: an fNIRS study. Brain Sci. 13:141. doi: 10.3390/brainsci13010141
Ma, Z. Z., Wu, J. J., Cao, Z., Hua, X. Y., Zheng, M. X., Xing, X. X., et al. (2024). Motor imagery-based brain–computer interface rehabilitation programs enhance upper extremity performance and cortical activation in stroke patients. J. Neuroeng. Rehabil. 21:91. doi: 10.1186/s12984-024-01387-w
Maier, M., Ballester, B. R., and Verschure, P. F. M. J. (2019). Principles of neurorehabilitation after stroke based on motor learning and brain plasticity mechanisms. Front. Syst. Neurosci. 13:74. doi: 10.3389/fnsys.2019.00074
Matarasso, A. K., Rieke, J. D., White, K., Yusufali, M. M., and Daly, J. J. (2021). Combined real-time fMRI and real time fNIRS brain computer interface (BCI): training of volitional wrist extension after stroke, a case series pilot study. PLoS One 16:e0250431. doi: 10.1371/journal.pone.0250431
Matsuo, M., Iso, N., Fujiwara, K., Moriuchi, T., Matsuda, D., Mitsunaga, W., et al. (2021). Comparison of cerebral activation between motor execution and motor imagery of self-feeding activity. Neural Regener. Res. 16:778. doi: 10.4103/1673-5374.295333
Mehler, D. M. A., Williams, A. N., Whittaker, J. R., Krause, F., Lührs, M., Kunas, S., et al. (2020). Graded fMRI neurofeedback training of motor imagery in middle cerebral artery stroke patients: a preregistered proof-of-concept study. Front. Hum. Neurosci. 14:226. doi: 10.3389/fnhum.2020.00226
Mihara, M., Fujimoto, H., Hattori, N., Otomune, H., Kajiyama, Y., Konaka, K., et al. (2021). Effect of neurofeedback facilitation on poststroke gait and balance recovery: a randomized controlled trial. Neurology 96, E2587–E2598. doi: 10.1212/WNL.0000000000011989
Moran, A., and O'Shea, H. (2020). Motor imagery practice and cognitive processes. Front. Psychol. 11:394. doi: 10.3389/fpsyg.2020.00394
Muci, B., Keser, I., Meric, A., and Karatas, G. K. (2022). What are the factors affecting dual-task gait performance in people after stroke?. Physiother. Theory Pract. 38, 621–628. doi: 10.1080/09593985.2020.1777603
Muller, C. O., Faity, G., Muthalib, M., Perrey, S., Dray, G., Xu, B., et al. (2024). Brain-movement relationship during upper-limb functional movements in chronic post-stroke patients. J. Neuroeng. Rehabil. 21:188. doi: 10.1186/s12984-024-01461-3
Ni, J., Jiang, W., Gong, X., Fan, Y., Qiu, H., Dou, J., et al. (2023). Effect of rTMS intervention on upper limb motor function after stroke: a study based on fNIRS. Front. Aging Neurosci. 14:1077218. doi: 10.3389/fnagi.2022.1077218
Noh, H. J., Lee, S. H., and Bang, D. H. (2019). Three-dimensional balance training using visual feedback on balance and walking ability in subacute stroke patients: a single-blinded randomized controlled pilot trial. J. Stroke Cerebrovasc. Dis. 28, 994–1000. doi: 10.1016/j.jstrokecerebrovasdis.2018.12.016
Nonnekes, J., Dibilio, V., Barthel, C., Solis-Escalante, T., Bloem, B. R., and Weerdesteyn, V. (2020). Understanding the dual-task costs of walking: a StartReact study. Exp. Brain Res. 238, 1359–1364. doi: 10.1007/s00221-020-05817-8
Nosaka, S., Imada, K., Saita, K., and Okamura, H. (2023). Prefrontal activation during dual-task seated stepping and walking performed by subacute stroke patients with hemiplegia. Front. Neurosci. 17:1169744. doi: 10.3389/fnins.2023.1169744
Ohzuno, T., and Usuda, S. (2019). Cognitive-motor interference in post-stroke individuals and healthy adults under different cognitive load and task prioritization conditions. J. Phys. Ther. Sci. 31, 255–260. doi: 10.1589/jpts.31.255
Ou, Z. T., Ding, Q., Yao, S. T., Zhang, L., Li, Y. W., Lan, Y., et al. (2024). Functional near-infrared spectroscopy evidence of cognitive-motor interference in different dual tasks. Eur. J. Neurosci. 59, 3045–3060. doi: 10.1111/ejn.16333
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., et al. (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 372:n71. doi: 10.1136/bmj.n71
Palidis, J. D., Gardiner, Z., Stephenson, A., Zhang, K., Boruff, J., and Fellows, L. K. (2025). The use of extrinsic performance feedback and reward to enhance upper limb motor behavior and recovery post-stroke: a scoping review. Neurorehabil. Neural Repair 39, 157–173. doi: 10.1177/15459683241298262
Pan, Y., Borragán, G., and Peigneux, P. (2019). Applications of functional near-infrared spectroscopy in fatigue, sleep deprivation, and social cognition. Brain Topogr. 32, 998–1012. doi: 10.1007/s10548-019-00740-w
Parker, S. M., Andreasen, S. C., Ricks, B., Kaipust, M. S., Zuniga, J., and Knarr, B. A. (2024). Comparison of brain activation and functional outcomes between physical and virtual reality box and block test: a case study. Disabil. Rehabil. Assist. Technol. 19, 273–280. doi: 10.1080/17483107.2022.2085334
Peng, Y., Zheng, Y., Yuan, Z., Guo, J., Fan, C., Li, C., et al. (2023). The characteristics of brain network in patient with post-stroke depression under cognitive task condition. Front. Neurosci. 17:1242543. doi: 10.3389/fnins.2023.1242543
Potts, C. A., Williamson, R. A., Jacob, J. D., Kantak, S. K., and Buxbaum, L. J. (2024). Reaching the cognitive-motor interface: effects of cognitive load on arm choice and motor performance after stroke. Exp. Brain Res. 242, 2785–2797. doi: 10.1007/s00221-024-06939-z
Qin, Y., Li, B., Wang, W., Shi, X., Peng, C., Wang, X., et al. (2025). ECA-FusionNet: a hybrid EEG-fNIRS signals network for MI classification. J. Neural Eng. 22. doi: 10.1088/1741-2552/adaf58
Qu, J., Du, Y., Jing, J., Wang, J., Bu, L., and Wang, Y. (2025). Short-term longitudinal study on brain network informatics of stroke patients under acupuncture and motor imagery intervention. IEEE J. Biomed. Health Inform. 29, 3356–3365. doi: 10.1109/JBHI.2025.3527074
Rendos, N. K., Zajac-Cox, L., Thomas, R., Sato, S., Eicholtz, S., and Kesar, T. M. (2021). Verbal feedback enhances motor learning during post-stroke gait retraining. Top. Stroke Rehabil. 28, 362–377. doi: 10.1080/10749357.2020.1818480
Rieke, J. D., Matarasso, A. K., Yusufali, M. M., Ravindran, A., Alcantara, J., White, K. D., et al. (2020). Development of a combined, sequential real-time fMRI and fNIRS neurofeedback system to enhance motor learning after stroke. J. Neurosci. Methods 341:108719. doi: 10.1016/j.jneumeth.2020.108719
Sakurada, T., Goto, A., Tetsuka, M., Nakajima, T., Morita, M., Yamamoto, S. I., et al. (2019). Prefrontal activity predicts individual differences in optimal attentional strategy for preventing motor performance decline: a functional near-infrared spectroscopy study. Neurophotonics 6:025012. doi: 10.1117/1.NPh.6.2.025012
Schroeter, M. L., Zysset, S., Kupka, T., Kruggel, F., and von Cramon, D. Y. (2002). Near-infrared spectroscopy can detect brain activity during a color–word matching stroop task in an event-related design. Hum. Brain Mapp. 17, 61–71. doi: 10.1002/hbm.10052
Schumacher, R., Halai, A. D., and Lambon Ralph, M. A. (2019). Assessing and mapping language, attention and executive multidimensional deficits in stroke aphasia. Brain 142, 3202–3216. doi: 10.1093/brain/awz258
Seitz, S., Schuster-Amft, C., Wandel, J., Bonati, L. H., Parmar, K., Gerth, H. U., et al. (2024). Effect of concurrent action observation, peripheral nerve stimulation and motor imagery on dexterity in patients after stroke: a pilot study. Sci. Rep. 14:14858. doi: 10.1038/s41598-024-65911-7
Shen, Q. Q., Hou, J. M., Xia, T., Zhang, J. Y., Wang, D. L., Yang, Y., et al. (2024). Exercise promotes brain health: a systematic review of fNIRS studies. Front. Psychol. 15:1327822. doi: 10.3389/fpsyg.2024.1327822
Shin, J., and Chung, Y. (2022). The effects of treadmill training with visual feedback and rhythmic auditory cue on gait and balance in chronic stroke patients: a randomized controlled trial. NeuroRehabilitation 51, 443–453. doi: 10.3233/NRE-220099
Shin, S., Lee, H. J., Chang, W. H., Ko, S. H., Shin, Y. I., and Kim, Y. H. (2022). A smart glove digital system promotes restoration of upper limb motor function and enhances cortical hemodynamic changes in subacute stroke patients with mild to moderate weakness: a randomized controlled trial. J. Clin. Med. 11:7343. doi: 10.3390/jcm11247343
Si, X., Li, S., Xiang, S., Yu, J., and Ming, D. (2021). Imagined speech increases the hemodynamic response and functional connectivity of the dorsal motor cortex. J. Neural Eng. 18. doi: 10.1088/1741-2552/ac25d9
Singh, S. (2024). Neuroplasticity and rehabilitation: harnessing brain plasticity for stroke recovery and functional improvement. Univers. Res. Rep. 11, 50–56. doi: 10.36676/urr.v11.i3.1287
Skau, S., Jonsdottir, I. H., Sjörs Dahlman, A., Johansson, B., and Kuhn, H. G. (2021). Exhaustion disorder and altered brain activity in frontal cortex detected with fNIRS. Stress 24, 64–75. doi: 10.1080/10253890.2020.1777972
Slim, K., Nini, E., Forestier, D., Kwiatkowski, F., Panis, Y., and Chipponi, J. (2003). Methodological index for non-randomized studies (minors): development and validation of a new instrument. ANZ J. Surg. 73, 712–716. doi: 10.1046/j.1445-2197.2003.02748.x
Stefaniak, J. D., Alyahya, R. S. W., and Lambon Ralph, M. A. (2021). Language networks in aphasia and health: a 1000 participant activation likelihood estimation meta-analysis. Neuroimage 233:117960. doi: 10.1016/j.neuroimage.2021.117960
Stephens, J. A., Mingils, S., and Orlandi, S. (2023). Evaluating dual task neurological costs with functional near-infrared spectroscopy: a preliminary report in healthy athletes. J. Integr. Neurosci. 22:133. doi: 10.31083/j.jin2205133
Strobach, T. (2020). The dual-task practice advantage: empirical evidence and cognitive mechanisms. Psychon. Bull. Rev. 27, 3–14. doi: 10.3758/s13423-019-01619-4
Strobach, T. (2024). Cognitive control and meta-control in dual-task coordination. Psychon. Bull. Rev. 31, 1445–1460. doi: 10.3758/s13423-023-02427-7
Su, W. C., Dashtestani, H., Miguel, H. O., Condy, E., Buckley, A., Park, S., et al. (2023). Simultaneous multimodal fNIRS-EEG recordings reveal new insights in neural activity during motor execution, observation, and imagery. Sci. Rep. 13:5151. doi: 10.1038/s41598-023-31609-5
Sun, R., Li, X., Zhu, Z., Li, T., Zhao, M., Mo, L., et al. (2022). Effects of dual-task training in patients with post-stroke cognitive impairment: a randomized controlled trial. Front. Neurol. 13:1027104. doi: 10.3389/fneur.2022.1027104
Sunwoo, J., Shah, P., Thuptimdang, W., Khaleel, M., Chalacheva, P., Kato, R. M., et al. (2023). Functional near-infrared spectroscopy-based prefrontal cortex oxygenation during working memory tasks in sickle cell disease. Neurophotonics 10:45004. doi: 10.1117/1.NPh.10.4.045004
Taguchi, J., Takami, A., and Makino, M. (2022). Changes in cerebral blood flow before, during, and after forward and backward walking in stroke patients trained using virtual reality walking videos with deliberately induced inaccuracies in walking speed estimations. J. Phys. Ther. Sci. 34, 668–672. doi: 10.1589/jpts.34.668
Takahashi, R., Fujita, K., Kobayashi, Y., Ogawa, T., Teranishi, M., and Kawamura, M. (2021). Effect of muscle fatigue on brain activity in healthy individuals. Brain Res. 1764:147469. doi: 10.1016/j.brainres.2021.147469
Tamashiro, H., Kinoshita, S., Okamoto, T., Urushidani, N., and Abo, M. (2019). Effect of baseline brain activity on response to low-frequency rTMS/intensive occupational therapy in poststroke patients with upper limb hemiparesis: a near-infrared spectroscopy study. Int. J. Neurosci. 129, 337–343. doi: 10.1080/00207454.2018.1536053
Tetsuka, M., Sakurada, T., Matsumoto, M., Nakajima, T., Morita, M., Fujimoto, S., et al. (2023). Higher prefrontal activity based on short-term neurofeedback training can prevent working memory decline in acute stroke. Front. Syst. Neurosci. 17:1130272. doi: 10.3389/fnsys.2023.1130272
Tsang, C. S. L., Wang, S., Miller, T., and Pang, M. Y. C. (2022). Degree and pattern of dual-task interference during walking vary with component tasks in people after stroke: a systematic review. J. Physiother. 68, 26–36. doi: 10.1016/j.jphys.2021.12.009
Udina, C., Avtzi, S., Durduran, T., Holtzer, R., Rosso, A. L., Castellano-Tejedor, C., et al. (2020). Functional near-infrared spectroscopy to study cerebral hemodynamics in older adults during cognitive and motor tasks: a review. Front. Aging Neurosci. 11:367. doi: 10.3389/fnagi.2019.00367
Varkanitsa, M., and Kiran, S. (2022). Understanding, facilitating and predicting aphasia recovery after rehabilitation. Int. J. Speech-Lang. Pathol. 24, 248–259. doi: 10.1080/17549507.2022.2075036
Villa-Berges, E., Laborda Soriano, A. A., Lucha-López, O., Tricas-Moreno, J. M., Hernández-Secorún, M., Gómez-Martínez, M., et al. (2023). Motor imagery and mental practice in the subacute and chronic phases in upper limb rehabilitation after stroke: a systematic review. Occup. Ther. Int. 2023:3752889. doi: 10.1155/2023/3752889
Wang, H., Xiong, X., Zhang, K., Wang, X., Sun, C., Zhu, B., et al. (2023). Motor network reorganization after motor imagery training in stroke patients with moderate to severe upper limb impairment. CNS Neurosci. Ther. 29, 619–632. doi: 10.1111/cns.14065
Wang, J., Jia, X., Song, J., Qin, Z., Cao, M., and Chen, J. (2025a). Motor cortex activation patterns in both hemispheres induced by motor imagery in patients with right- and left-sided cerebral infarction: an fNIRS study. Eur. J. Neurosci. 61:e70079. doi: 10.1111/ejn.70079
Wang, J., Zhou, J., Zhu, J., Sheng, J., Jiang, R., and Zhang, X. (2025b). Brain remodeling in stroke patients: a comprehensive review of mechanistic and neuroimaging studies. Behav. Brain Res. 486:115548. doi: 10.1016/j.bbr.2025.115548
Wang, L., Gao, F., Dai, Y., Wang, Z., Liang, F., Wu, J., et al. (2023). Transcutaneous auricular vagus nerve stimulation on upper limb motor function with stroke: a functional near-infrared spectroscopy pilot study. Front. Neurosci. 17:1297887. doi: 10.3389/fnins.2023.1297887
Wang, L., Li, Y., Liu, R., Li, H., Wang, L., Yuan, Y., et al. (2024). The effect and mechanism of motor imagery based on action observation treatment on dysphagia in wallenberg syndrome: a randomized controlled trial. Eur. J. Phys. Rehabil. Med. 60, 938–948. doi: 10.23736/S1973-9087.24.08471-5
Wang, M. H., Wang, Y. X., Xie, M., Chen, L. Y., He, M. F., Lin, F., et al. (2024). Transcutaneous auricular vagus nerve stimulation with task-oriented training improves upper extremity function in patients with subacute stroke: a randomized clinical trial. Front. Neurosci. 18. doi: 10.3389/fnins.2024.1346634
Wang, Q., Dai, W., Xu, S., Zhu, S., Sui, Y., Kan, C., et al. (2023). Brain activation of the PFC during dual-task walking in stroke patients: a systematic review and meta-analysis of functional near-infrared spectroscopy studies. Front. Neurosci. 17:1111274. doi: 10.3389/fnins.2023.1111274
Wang, Z., Cao, C., Chen, L., Gu, B., Liu, S., Xu, M., et al. (2022). Multimodal neural response and effect assessment during a BCI-based neurofeedback training after stroke. Front. Neurosci. 16:884420. doi: 10.3389/fnins.2022.884420
Wang, Z., Yang, L., Wang, M., Zhou, Y., Chen, L., Gu, B., et al. (2023). Motor imagery and action observation induced electroencephalographic activations to guide subject-specific training paradigm: a pilot study. IEEE Trans. Neural Syst. Rehabil. Eng. 31, 2457–2467. doi: 10.1109/TNSRE.2023.3275572
Wen, X., Peng, J., Zhu, Y., Bao, X., Wan, Z., Hu, R., et al. (2023). Hemodynamic signal changes and functional connectivity in acute stroke patients with dysphagia during volitional swallowing: a pilot study. Med. Phys. 50, 5166–5175. doi: 10.1002/mp.16535
Wriessnegger, S. C., Kirchmeyr, D., Bauernfeind, G., and Müller-Putz, G. R. (2017). Force related hemodynamic responses during execution and imagery of a hand grip task: a functional near infrared spectroscopy study. Brain Cogn. 117, 108–116. doi: 10.1016/j.bandc.2017.06.010
Xia, W., Dai, R., Xu, X., Huai, B., Bai, Z., Zhang, J., et al. (2022). Cortical mapping of active and passive upper limb training in stroke patients and healthy people: a functional near-infrared spectroscopy study. Brain Res. 1788:147935. doi: 10.1016/j.brainres.2022.147935
Xia, Y., Tang, X., Hu, R., Liu, J., Zhang, Q., Tian, S., et al. (2022). Cerebellum-cerebrum paired target magnetic stimulation on balance function and brain network of patients with stroke: a functional near-infrared spectroscopy pilot study. Front. Neurol. 13:1071328. doi: 10.3389/fneur.2022.1071328
Xie, H., Jing, J., Ma, Y., Song, Y., Yin, J., Xu, G., et al. (2022a). Effects of simultaneous use of m-NMES and language training on brain functional connectivity in stroke patients with aphasia: a randomized controlled clinical trial. Front. Aging Neurosci. 14:965486. doi: 10.3389/fnagi.2022.965486
Xie, H., Li, X., Huang, W., Yin, J., Luo, C., Li, Z., et al. (2022b). Effects of robot-assisted task-oriented upper limb motor training on neuroplasticity in stroke patients with different degrees of motor dysfunction: a neuroimaging motor evaluation index. Front. Neurosci. 16:957972. doi: 10.3389/fnins.2022.957972
Xu, G., Huo, C., Yin, J., Li, W., Xie, H., Li, X., et al. (2022). Effective brain network analysis in unilateral and bilateral upper limb exercise training in subjects with stroke. Med. Phys. 49, 3333–3346. doi: 10.1002/mp.15570
Xu, G., Huo, C., Yin, J., Zhong, Y., Sun, G., Fan, Y., et al. (2023). Test-retest reliability of fNIRS in resting-state cortical activity and brain network assessment in stroke patients. Biomed. Opt. Express 14, 4217–4236. doi: 10.1364/BOE.491610
Xu, R., Zhu, G. Y., Zhu, J., Wang, Y., Xing, X. X., Chen, L. Y., et al. (2022). Using hebbian-type stimulation to rescue arm function after stroke: study protocol for a randomized clinical trial. Front. Neural Circuits 15:789095. doi: 10.3389/fncir.2021.789095
Yang, C., Zhang, T., Huang, K., Xiong, M., Liu, H., Wang, P., et al. (2022). Increased both cortical activation and functional connectivity after transcranial direct current stimulation in patients with post-stroke: a functional near-infrared spectroscopy study. Front. Psychiatry 13:1046849. doi: 10.3389/fpsyt.2022.1046849
Yang, Z., Ye, L., Yang, L., Lu, Q., Yu, A., and Bai, D. (2024). Early screening of post-stroke fall risk: a simultaneous multimodal fNIRs-EMG study. CNS Neurosci. Ther. 30:e70041. doi: 10.1111/cns.70041
Ye, S., Tao, L., Gong, S., Ma, Y., Wu, J., Li, W., et al. (2024). Upper limb motor assessment for stroke with force, muscle activation and interhemispheric balance indices based on sEMG and fNIRS. Front. Neurol. 15:1337230. doi: 10.3389/fneur.2024.1337230
Yu, H., Zheng, B., Zhang, Y., Chu, M., Shu, X., Wang, X., et al. (2024). Activation changes in patients with post-stroke cognitive impairment receiving intermittent theta burst stimulation: a functional near-infrared spectroscopy study. NeuroRehabilitation 54, 677–690. doi: 10.3233/NRE-240068
Yu, J., Zhang, X., Yang, J., Wang, Z., Zhao, H., Yuan, X., et al. (2022). A functional near-infrared spectroscopy study of the effects of video game-based bilateral upper limb training on brain cortical activation and functional connectivity. Exp. Gerontol. 169:111962. doi: 10.1016/j.exger.2022.111962
Yu, Y., Shen, X., Hong, Y., and Wang, F. (2024). Characteristic brain functional activation and connectivity during actual and imaginary right-handed grasp. Brain Res. 1844:149141. doi: 10.1016/j.brainres.2024.149141
Yuan, Z., Peng, Y., Wang, L., Song, S., Chen, S., Yang, L., et al. (2021). Effect of BCI-controlled pedaling training system with multiple modalities of feedback on motor and cognitive function rehabilitation of early subacute stroke patients. IEEE Trans. Neural Syst. Rehabil. Eng. 29, 2569–2577. doi: 10.1109/TNSRE.2021.3132944
Yuan, Z., Xu, W., Bao, J., Gao, H., Li, W., Peng, Y., et al. (2022). Task-state cortical motor network characteristics by functional near-infrared spectroscopy in subacute stroke show hemispheric dominance. Front. Aging Neurosci. 14:932318. doi: 10.3389/fnagi.2022.932318
Zhang, K., Sun, X., Yu, C. L., Eggleston, R. L., Marks, R. A., Nickerson, N., et al. (2023). Phonological and morphological literacy skills in English and Chinese: a cross-linguistic neuroimaging comparison of chinese-english bilingual and monolingual english children. Hum. Brain Mapp. 44, 4812–4829. doi: 10.1002/hbm.26419
Zhang, Y. F., Lasfargues-Delannoy, A., and Berry, I. (2022). Adaptation of stimulation duration to enhance auditory response in fNIRS block design. Hear. Res. 424:108593. doi: 10.1016/j.heares.2022.108593
Zhao, J. L., Chen, P. M., Zhang, T., Xie, H., Xiao, W. W., Ng, S. S. M., et al. (2024). Characteristics of central cortex and upper-limb flexors synchrony oxygenation during grasping in people with stroke: a controlled trial study protocol. Front. Hum. Neurosci. 18:1409148. doi: 10.3389/fnhum.2024.1409148
Zhao, W., Makowski, C., Hagler, D. J., Garavan, H. P., Thompson, W. K., Greene, D. J., et al. (2023). Task fMRI paradigms may capture more behaviorally relevant information than resting-state functional connectivity. Neuroimage 270:119946. doi: 10.1016/j.neuroimage.2023.119946
Zou, J., Yin, Y., Lin, Z., and Gong, Y. (2023). The analysis of brain functional connectivity of post-stroke cognitive impairment patients: an fNIRS study. Front. Neurosci. 17:1168773. doi: 10.3389/fnins.2023.1168773
Keywords: brain region, fNIRS, paradigm, rehabilitation, stroke
Citation: Huang Y, Zhan X, Zeng H, Li S, Shi J, Cui Z, Fan Q, Li B, Sui Y, Liang F and Song Z (2026) A systematic review of functional near-infrared spectroscopy-based task paradigms in stroke rehabilitation. Front. Hum. Neurosci. 19:1633142. doi: 10.3389/fnhum.2025.1633142
Received: 22 May 2025; Revised: 16 December 2025; Accepted: 24 December 2025;
Published: 16 January 2026.
Edited by:
Filippo Brighina, University of Palermo, ItalyReviewed by:
Rupert Ortner, g.tec medical engineering GmbH, AustriaChuan Guo, Nanjing Medical University, China
Copyright © 2026 Huang, Zhan, Zeng, Li, Shi, Cui, Fan, Li, Sui, Liang and Song. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Fengyan Liang, ZnlsaWFuZ0BoYWluYW51LmVkdS5jbg==; Zhenhua Song, YTE5NzQ3ODEwMTBAMTYzLmNvbQ==
†These authors have contributed equally to this work and share first authorship