Shohei Tsuchimoto
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Shohei Tsuchimoto has a total of 59 co-authors. Here are the top co-authors by number of shared publications:
- Junichi Ushiba
Keio University
4 shared publications Nobuaki Mizuguchi
The Japan Society for the Promotion of Science, Tokyo, Japan; Department of Biosciences and informatics, Faculty of Science and Technology, Keio University, 3-14-1 Hiyoshi, Kouhoku-ku, Yokohama, Kanagawa 223-8522, Japan
4 shared publicationsMeigen Liu
Department of Rehabilitation Medicine, School of Medicine, Keio University, Tokyo, Japan
3 shared publications- Nobuaki Mizuguchi
Juntendo Administration for Sports, Health and Medical Sciences, Juntendo University, Japan
3 shared publications - Izumi Kondo
National Center for Geriatrics and Gerontology, Obu, Japan
2 shared publications Junichi Ushiba
Department of Biosciences and Informatics, Faculty of Science and Technology, Keio University; and Keio Institute of Pure and Applied Sciences, Kanagawa, Japan
2 shared publications
Latest Publications
- Open Access
- 13 November 2026
- Published by MIT Press in Imaging Neuroscience
Beyond shared input: verbal communication transmits partner-specific perceptual representations across brains
Language allows us to communicate what we perceive, but the degree to which such exchanges produce shared perceptual representations between interlocutors is unclear. In particular, whether verbal communication gives rise to interaction-specific neural alignment that reflects the speaker's perceptual experience remains untested. Here we combined fMRI hyperscanning with inter-brain spatial pattern similarity analyses to address this question. Senders viewed face images and described them to receivers via a real-time audio link while both underwent fMRI. Each trial comprised an utter–listen period, in which the sender described a face while the receiver listened, followed by a silent view–imagery period, in which the sender re-viewed the face while the receiver imagined it. Sender-receiver neural similarity was assessed by comparing true communicative dyads with pseudo pairs matched for stimulus content and timing. True pairs exhibited greater describe-to-imagery spatial pattern similarity than pseudo pairs across a left-lateralized frontoparietal network, indicating that the receiver's neural representation partially reflected the sender's verbal encoding in a pair-specific manner (describe-to-imagery similarity). Pair-specific similarity was also observed between patterns from the post-description view–imagery period (view-to-imagery similarity). Across both analyses, effects were concentrated in higher-order integrative and semantic regions, with limited evidence in early visual and face-selective cortex. Furthermore, sender–receiver similarity increased more steeply over the course of each description in true than in pseudo pairs (within-trial temporal accumulation). These findings provide neural evidence that verbal communication is associated with partner-specific representational alignment in the receiver's brain, emerging at an integrated, conceptual level and building over the course of each communicative exchange.
- 13 November 2026
- Published by MIT Press in Imaging Neuroscience
Open AccessBeyond shared input: verbal communication transmits partner-specific perceptual representations across brains
Language allows us to communicate what we perceive, but the degree to which such exchanges produce shared perceptual representations between interlocutors is unclear. In particular, whether verbal communication gives rise to interaction-specific neural alignment that reflects the speaker's perceptual experience remains untested. Here we combined fMRI hyperscanning with inter-brain spatial pattern similarity analyses to address this question. Senders viewed face images and described them to receivers via a real-time audio link while both underwent fMRI. Each trial comprised an utter–listen period, in which the sender described a face while the receiver listened, followed by a silent view–imagery period, in which the sender re-viewed the face while the receiver imagined it. Sender-receiver neural similarity was assessed by comparing true communicative dyads with pseudo pairs matched for stimulus content and timing. True pairs exhibited greater describe-to-imagery spatial pattern similarity than pseudo pairs across a left-lateralized frontoparietal network, indicating that the receiver's neural representation partially reflected the sender's verbal encoding in a pair-specific manner (describe-to-imagery similarity). Pair-specific similarity was also observed between patterns from the post-description view–imagery period (view-to-imagery similarity). Across both analyses, effects were concentrated in higher-order integrative and semantic regions, with limited evidence in early visual and face-selective cortex. Furthermore, sender–receiver similarity increased more steeply over the course of each description in true than in pseudo pairs (within-trial temporal accumulation). These findings provide neural evidence that verbal communication is associated with partner-specific representational alignment in the receiver's brain, emerging at an integrated, conceptual level and building over the course of each communicative exchange.
- Open Access
- 26 August 2026
- Published by Elsevier in Geriatric Nursing: Asia
- Vol. 2
An observational time-motion and heart-rate study of nursing care staff workload
- 26 August 2026
- Published by Elsevier in Geriatric Nursing: Asia
- Vol. 2
Open AccessAn observational time-motion and heart-rate study of nursing care staff workload
- Open Access
- 25 June 2026
- Published by Elsevier in SSRN Electronic Journal
Pretraining Objective Shapes Cross-Category Generalization in Affective Image Prediction: A Geometric Analysis of Vision Transformer Encoders
The geometry of representations learned by deep neural networks is shaped jointly by architecture and pretraining objective, yet disentangling these two factors remains difficult. Here we isolate the contribution of pretraining objective by comparing two Vision Transformers from the same backbone family but trained under different objectives: language–image contrastive learning (CLIP) and ImageNet-21k classification. Using continuous Valence–Arousal prediction on the OASIS dataset as a probe of representational quality, we evaluated frozen features under Leave-One-Theme-Out and Leave-One-Category-Out cross-validation, the latter requiring extrapolation to entirely unseen semantic categories. The contrastively pretrained encoder generalized substantially better than the classification-pretrained encoder under both protocols, with the gap widening sharply when held-out categories required cross-category generalization. To characterize why the two representations differ, we developed a geometric analysis of prediction errors, treating per-image errors as vectors in the affective plane and quantifying their spatial structure via weighted phase-locking, trajectory-based occupancy entropy, and effective dimensionality. The classification-pretrained representation collapsed errors into a small number of attractor regions with a strong center-ward pull, whereas the language-aligned representation distributed errors broadly across the affective space. Layer-wise linear probing further revealed that affective information was distributed across depth in the contrastive encoder but increasingly concentrated in deeper layers of the classification encoder, mirroring the texture-bias and category-anchored statistics characteristic of ImageNet-trained representations. These results provide a representation-geometric account of how the choice of pretraining objective, holding architecture constant, determines whether learned features generalize across semantic boundaries or remain confined to category-bound visual regularities.
- 25 June 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessPretraining Objective Shapes Cross-Category Generalization in Affective Image Prediction: A Geometric Analysis of Vision Transformer Encoders
The geometry of representations learned by deep neural networks is shaped jointly by architecture and pretraining objective, yet disentangling these two factors remains difficult. Here we isolate the contribution of pretraining objective by comparing two Vision Transformers from the same backbone family but trained under different objectives: language–image contrastive learning (CLIP) and ImageNet-21k classification. Using continuous Valence–Arousal prediction on the OASIS dataset as a probe of representational quality, we evaluated frozen features under Leave-One-Theme-Out and Leave-One-Category-Out cross-validation, the latter requiring extrapolation to entirely unseen semantic categories. The contrastively pretrained encoder generalized substantially better than the classification-pretrained encoder under both protocols, with the gap widening sharply when held-out categories required cross-category generalization. To characterize why the two representations differ, we developed a geometric analysis of prediction errors, treating per-image errors as vectors in the affective plane and quantifying their spatial structure via weighted phase-locking, trajectory-based occupancy entropy, and effective dimensionality. The classification-pretrained representation collapsed errors into a small number of attractor regions with a strong center-ward pull, whereas the language-aligned representation distributed errors broadly across the affective space. Layer-wise linear probing further revealed that affective information was distributed across depth in the contrastive encoder but increasingly concentrated in deeper layers of the classification encoder, mirroring the texture-bias and category-anchored statistics characteristic of ImageNet-trained representations. These results provide a representation-geometric account of how the choice of pretraining objective, holding architecture constant, determines whether learned features generalize across semantic boundaries or remain confined to category-bound visual regularities.
- Open Access
- 1 January 2026
- Published by Wiley in Neural Plasticity
- Vol. 2026 (1) , e8819722
Functional–Structural Plasticity Associated With the Duration of Sports Participation in Lower Limb Amputees: A Multimodal Neuroimaging Study
Structural and functional brain reorganization can occur after long‐term physical activity and lower limb amputation (LLA). A previous study suggested that activation of the primary motor cortex (M1) ipsilateral to the amputated leg during rectus femoris contraction is associated with the amount of sports participation in individuals with LLA. However, the structural basis of ipsilateral M1 activation remains unclear. The aim of this study was to investigate whether ipsilateral M1 activation is associated with white matter microstructure in descending motor pathways. We hypothesized that ipsilateral M1 activation would be related to microstructural properties of the corticoreticular tract (CRT) rather than the corticospinal tract (CST). Twenty‐three individuals with LLA who had participated in sports for varying durations underwent functional magnetic resonance imaging (fMRI) and diffusion‐weighted imaging. During fMRI, percent signal change (PSC) during rectus femoris contraction in the amputated leg was quantified. White matter microstructure was assessed using fixel‐based analysis (FBA). Ipsilateral M1 PSC during contraction of the amputated leg was positively correlated with fiber cross‐section (FC) in the CRT within the same hemisphere, whereas no significant correlation was observed for the CST. No significant correlations were found in control analyses using PSCs during contraction of the rectus femoris in the nonamputated leg, during motor imagery of ankle movements of the amputated leg, or in the visual cortex. These findings indicate a task‐ and hemisphere‐specific functional–structural association between ipsilateral M1 activation and the CRT. Although FC in the CRT was not directly correlated with the amount of sports participation, it may constrain the extent of ipsilateral M1 functional reorganization associated with long‐term sports involvement. Our results highlight a potential role of the CRT, in addition to the CST, in motor plasticity and rehabilitation in individuals with LLA.
- 1 January 2026
- Published by Wiley in Neural Plasticity
- Vol. 2026 (1) , e8819722
Open AccessFunctional–Structural Plasticity Associated With the Duration of Sports Participation in Lower Limb Amputees: A Multimodal Neuroimaging Study
Structural and functional brain reorganization can occur after long‐term physical activity and lower limb amputation (LLA). A previous study suggested that activation of the primary motor cortex (M1) ipsilateral to the amputated leg during rectus femoris contraction is associated with the amount of sports participation in individuals with LLA. However, the structural basis of ipsilateral M1 activation remains unclear. The aim of this study was to investigate whether ipsilateral M1 activation is associated with white matter microstructure in descending motor pathways. We hypothesized that ipsilateral M1 activation would be related to microstructural properties of the corticoreticular tract (CRT) rather than the corticospinal tract (CST). Twenty‐three individuals with LLA who had participated in sports for varying durations underwent functional magnetic resonance imaging (fMRI) and diffusion‐weighted imaging. During fMRI, percent signal change (PSC) during rectus femoris contraction in the amputated leg was quantified. White matter microstructure was assessed using fixel‐based analysis (FBA). Ipsilateral M1 PSC during contraction of the amputated leg was positively correlated with fiber cross‐section (FC) in the CRT within the same hemisphere, whereas no significant correlation was observed for the CST. No significant correlations were found in control analyses using PSCs during contraction of the rectus femoris in the nonamputated leg, during motor imagery of ankle movements of the amputated leg, or in the visual cortex. These findings indicate a task‐ and hemisphere‐specific functional–structural association between ipsilateral M1 activation and the CRT. Although FC in the CRT was not directly correlated with the amount of sports participation, it may constrain the extent of ipsilateral M1 functional reorganization associated with long‐term sports involvement. Our results highlight a potential role of the CRT, in addition to the CST, in motor plasticity and rehabilitation in individuals with LLA.
- Open Access
- 30 September 2022
- Published by Taylor & Francis in Disability and Rehabilitation: Assistive Technology
- Vol. 19 (3) , 841-850
Investigating proficiency using a lift-type transfer support device for effective care: comparison of skilled and unskilled nursing homes
The purpose of the study was to investigate whether the sustained use of the "Hug," a "hugging" type robotic transfer support device, could increase the level of quality of care.The effect of profi...
- 30 September 2022
- Published by Taylor & Francis in Disability and Rehabilitation: Assistive Technology
- Vol. 19 (3) , 841-850
Open AccessInvestigating proficiency using a lift-type transfer support device for effective care: comparison of skilled and unskilled nursing homes
The purpose of the study was to investigate whether the sustained use of the "Hug," a "hugging" type robotic transfer support device, could increase the level of quality of care.The effect of profi...
- Open Access
- 1 December 2020
- Published by Elsevier in NeuroImage
- Vol. 223, 117298
Neurofeedback of scalp bi-hemispheric EEG sensorimotor rhythm guides hemispheric activation of sensorimotor cortex in the targeted hemisphere
- 1 December 2020
- Published by Elsevier in NeuroImage
- Vol. 223, 117298
Open AccessNeurofeedback of scalp bi-hemispheric EEG sensorimotor rhythm guides hemispheric activation of sensorimotor cortex in the targeted hemisphere
- Open Access
- 1 November 2020
- Published by Elsevier in NeuroImage
- Vol. 222, 117249
Scalp electroencephalograms over ipsilateral sensorimotor cortex reflect contraction patterns of unilateral finger muscles
- 1 November 2020
- Published by Elsevier in NeuroImage
- Vol. 222, 117249
Open AccessScalp electroencephalograms over ipsilateral sensorimotor cortex reflect contraction patterns of unilateral finger muscles
- 12 February 2020
- Published by Oxford University Press (OUP) in Physical Therapy
- Vol. 100 (5) , 870-879
Depth Sensor–Based Assessment of Reachable Work Space for Visualizing and Quantifying Paretic Upper Extremity Motor Function in People With Stroke
Background: Quantitative evaluation of upper extremity (UE) motor function is important in people with hemiparetic stroke. A depth sensor–based assessment of reachable work space (RWS) was applied to visualize and quantify paretic UE motor function.Objective: The objectives of this study were to examine the characteristics of RWS and to assess its validity, reliability, measurement error, and responsiveness in people with hemiparetic stroke.Design: This was a descriptive, repeated-measures, observational study.Methods: Fifty-eight people with stroke participated. RWS was assessed on both paretic and nonparetic UEs, and the RWS ratio was determined by dividing the RWS of the paretic UE by that of the nonparetic UE. The concurrent validity of the RWS was determined by examining the relationship with the Fugl-Meyer Assessment UE motor score. Test-retest reproducibility was examined in 40 participants. Responsiveness was determined by examining the RWS results before and after 3 weeks of intensive training of the paretic UE in 32 participants.Results: The lower area of RWS bordering shoulder was significantly larger than the upper area, and the medial-lower area of RWS bordering shoulder was significantly larger than the lateral-lower area. The RWS ratio was highly correlated with the Fugl-Meyer Assessment UE motor score (r = 0.81). The RWS ratio showed good intrarater relative reliability (intraclass correlation coefficient = 0.94) and no fixed or proportional bias. The minimal detectable change of the RWS ratio was 16.6. The responsiveness of the RWS ratio was large (standardized response mean = 0.83).Limitations: Interexaminer reliability was not assessed.Conclusions: The RWS assessment showed sufficient validity, reliability, and responsiveness in people with hemiparetic stroke. A depth sensor–based RWS evaluation is useful for visualizing and quantifying paretic UE motor function in the clinical setting.
- 12 February 2020
- Published by Oxford University Press (OUP) in Physical Therapy
- Vol. 100 (5) , 870-879
Depth Sensor–Based Assessment of Reachable Work Space for Visualizing and Quantifying Paretic Upper Extremity Motor Function in People With Stroke
Background: Quantitative evaluation of upper extremity (UE) motor function is important in people with hemiparetic stroke. A depth sensor–based assessment of reachable work space (RWS) was applied to visualize and quantify paretic UE motor function.Objective: The objectives of this study were to examine the characteristics of RWS and to assess its validity, reliability, measurement error, and responsiveness in people with hemiparetic stroke.Design: This was a descriptive, repeated-measures, observational study.Methods: Fifty-eight people with stroke participated. RWS was assessed on both paretic and nonparetic UEs, and the RWS ratio was determined by dividing the RWS of the paretic UE by that of the nonparetic UE. The concurrent validity of the RWS was determined by examining the relationship with the Fugl-Meyer Assessment UE motor score. Test-retest reproducibility was examined in 40 participants. Responsiveness was determined by examining the RWS results before and after 3 weeks of intensive training of the paretic UE in 32 participants.Results: The lower area of RWS bordering shoulder was significantly larger than the upper area, and the medial-lower area of RWS bordering shoulder was significantly larger than the lateral-lower area. The RWS ratio was highly correlated with the Fugl-Meyer Assessment UE motor score (r = 0.81). The RWS ratio showed good intrarater relative reliability (intraclass correlation coefficient = 0.94) and no fixed or proportional bias. The minimal detectable change of the RWS ratio was 16.6. The responsiveness of the RWS ratio was large (standardized response mean = 0.83).Limitations: Interexaminer reliability was not assessed.Conclusions: The RWS assessment showed sufficient validity, reliability, and responsiveness in people with hemiparetic stroke. A depth sensor–based RWS evaluation is useful for visualizing and quantifying paretic UE motor function in the clinical setting.
- 1 September 2019
- Published by Elsevier in Neuroscience
- Vol. 416, 109-125
Sensorimotor Connectivity after Motor Exercise with Neurofeedback in Post-Stroke Patients with Hemiplegia
- 1 September 2019
- Published by Elsevier in Neuroscience
- Vol. 416, 109-125
Sensorimotor Connectivity after Motor Exercise with Neurofeedback in Post-Stroke Patients with Hemiplegia