TATomoyuki Akiyama

Tomoyuki Akiyama

*This is an algorithmically generated author record based on publicly available information.

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


Tomoyuki Akiyama has a total of 176 co-authors. Here are the top co-authors by number of shared publications:

  • Katsuhiro Kobayashi

    Department of Pediatrics (Child Neurology), Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama, Japan

    19 shared publications
  • Harumi Yoshinaga

    Department of Child Neurology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama, Japan

    9 shared publications
  • Takashi Shibata

    Department of Pediatric Neurology Okayama University Hospital Okayama Japan

    8 shared publications
  • Mari Akiyama

    Department of Pediatric Neurology Okayama University Hospital Okayama Japan

    7 shared publications
  • Hiroki Tsuchiya

    Department of Pediatric Neurology Okayama University Hospital Okayama Japan

    6 shared publications
  • Akira Furui

    Graduate School of Advanced Science and Engineering, Hiroshima University, Higashihiroshima, Japan

    5 shared publications

Latest Publications

  • Open Access

    EEG-Based Cross-Patient Epileptic Seizure Detection via State-Conditional Domain Adversarial Training and Temporal Modeling

    by Rina Tazaki, Tomoyuki Akiyama, Akira Furui

    Automated epileptic seizure detection from scalp electroencephalogram (EEG) remains challenging due to significant inter-patient variability. While domain adversarial training has emerged as a promising approach for patient-invariant features, existing methods typically rely on convolutional neural network (CNN)-based models that process each EEG window independently, failing to capture temporal progression patterns of seizures. Moreover, conventional domain adversarial training enforces patient invariance uniformly without distinguishing between seizure and non-seizure states, despite inter-patient variability differing across these states. To address these limitations, we propose a domain generalization framework that integrates state-conditional domain adversarial training with temporal modeling via a CNN and bidirectional long short-term memory (BiLSTM), using only source patient data without any target patient enrollment. The two objectives are decoupled into separate stages to avoid conflicting optimization. In the first stage, the CNN extracts state-adaptive patient-invariant features by conditional domain adversarial training. In the second stage, the BiLSTM captures temporal dependencies of these invariant features to model seizure evolution.We evaluated our method using leave-one-patient-out cross-validation on two focal epilepsy datasets: 14 patients from the Siena Scalp EEG Database (long-term recordings) and 20 from a clinical dataset (short-term recordings). The results demonstrate that both conditional domain adversarial training and temporal modeling contribute to performance improvement, with the benefit of state-adaptive invariant feature extraction particularly pronounced in long-term recordings containing diverse background EEG activities. This approach enables immediate application to unseen patients without any enrollment period, providing a clinically practical solution for automated seizure monitoring.

  • Open Access

    EEG-Based Inter-Patient Epileptic Seizure Detection Combining Domain Adversarial Training with CNN-BiLSTM Network

    by Rina Tazaki, Tomoyuki Akiyama, Akira Furui

    Automated epileptic seizure detection from electroencephalogram (EEG) remains challenging due to significant individual differences in EEG patterns across patients. While existing studies achieve high accuracy with patient-specific approaches, they face difficulties in generalizing to new patients. To address this, we propose a detection framework combining domain adversarial training with a convolutional neural network (CNN) and a bidirectional long short-term memory (BiLSTM). First, the CNN extracts local patient-invariant features through domain adversarial training, which optimizes seizure detection accuracy while minimizing patient-specific characteristics. Then, the BiLSTM captures temporal dependencies in the extracted features to model seizure evolution patterns. Evaluation using EEG recordings from 20 patients with focal epilepsy demonstrated superior performance over non-adversarial methods, achieving high detection accuracy across different patients. The integration of adversarial training with temporal modeling enables robust cross-patient seizure detection.


  • Vitamin B6 Status in Hypophosphatasia: Association With Clinical Severity, Diagnostic Utility, and Effects on Vitamin B6 Metabolism by Supplementation and Enzyme Replacement Therapy

    by Tomoyuki Akiyama, Takuo Kubota, Kei Murayama, Makoto Fujiwara, Yasuhisa Ohata, Keiichi Ozono, Makiko Tajika, Keiko Ichimoto, Yohei Sugiyama, Mari Akiyama, et al.

    This study reports the concentrations of major vitamin B6 (VB6) vitamers (pyridoxal 5′‐phospate [PLP], pyridoxal [PL], and 4‐pyridoxic acid [PA]) in over 100 hypophosphatasia (HPP) cases to explore clues for clinical severity, to improve diagnostic sensitivity, and to examine VB6 changes by supplementation and enzyme replacement therapy. Serum samples were collected from HPP and non‐HPP patients. When available, cerebrospinal fluid (CSF) samples were also analyzed to determine PLP, PL, and PA concentrations. Serum PLP concentrations, PLP‐to‐PL ratios, and PLP‐to‐PA ratios were higher in untreated HPP patients compared to non‐HPP patients and reflected clinical severity. Perinatal severe HPP showed lower PL concentrations than perinatal benign HPP and the lowest PA concentrations among all clinical forms. Combining PLP‐to‐PL ratios with PLP concentrations improved diagnostic sensitivity. Under VB6 therapy, PLP concentrations and PLP‐to‐PL ratios remained higher in HPP patients than in non‐HPP patients. VB6 changes produced by ERT were most clearly reflected by the reduction of PLP‐to‐PL ratios. CSF PLP concentrations and PLP‐to‐PL ratios were higher in untreated HPP patients than in non‐HPP patients, while CSF PL concentrations were higher in patients on ERT compared to those not on ERT. Co‐evaluation of serum PLP, PL, and PA concentrations may help understand clinical severity, can improve diagnostic sensitivity, and can demonstrate the effect of ERT on VB6 metabolism more effectively than the assay of PLP concentrations alone. Altered VB6 status in the CSF suggests that tissue‐nonspecific alkaline phosphatase plays a key role in VB6 transport from peripheral blood to the brain via the blood–CSF barrier.


  • Open Access

    Investigation of the relationship between 0.5–1200 Hz signal characteristics of cortical high-frequency oscillations and epileptogenicity through multivariate analysis

    by Takashi Shibata, Hiroki Tsuchiya, Mari Akiyama, Tomoyuki Akiyama, Masao Matsuhashi, Katsuhiro Kobayashi

  • Open Access

    Epileptic Seizure Detection Using a Recurrent Neural Network With Temporal Features Derived From a Scale Mixture EEG Model

    by Akira Furui, Ryota Onishi, Tomoyuki Akiyama, Toshio Tsuji

    Automated detection of epileptic seizures from scalp Electroencephalogram (EEG) is crucial for improving epilepsy diagnosis and management. This paper presents an automated inter-patient epileptic seizure detection method using multichannel EEG signals. The proposed method performs both feature extraction and seizure detection based on a scale mixture-based stochastic EEG model and a recurrent neural network, respectively. Specifically, the stochastic model that can consider uncertainties in the EEG amplitude is first fitted to a specific frequency band of EEG, thereby extracting relevant features of the seizure. Then, a recurrent neural network-based recognition architecture learns the temporal evolution of these features. We evaluated our method using EEG data from 20 patients with focal epilepsy, conducting comprehensive assessments including ablation studies on classifiers and features. The results demonstrate that our approach outperforms static classifiers and existing feature sets, achieving high sensitivity while maintaining acceptable specificity. Furthermore, our feature set showed efficacy both independently and as a complement to existing features, indicating its robustness in seizure detection tasks. These findings reveal that learning the temporal evolution of the stochastic fluctuation and amplitude information of EEG extracted using a stochastic model enables highly accurate seizure detection, potentially advancing automated epilepsy diagnosis in clinical settings.


  • Exploration of urine metabolic biomarkers for new-onset, untreated pediatric epilepsy: A gas and liquid chromatography mass spectrometry-based metabolomics study

    by Tomoyuki Akiyama, Daisuke Saigusa, Takushi Inoue, Chiho Tokorodani, Mari Akiyama, Rie Michiue, Atsushi Mori, Eiji Hishinuma, Naomi Matsukawa, Takashi Shibata, et al.


  • A female patient with adolescent-onset progressive myoclonus epilepsy carrying a truncating MECP2 mutation

    by Mari Akiyama, Tomoyuki Akiyama, Hirotomo Saitsu, Yukie Tokioka, Rie Tsukahara, Hiroki Tsuchiya, Takashi Shibata, Katsuhiro Kobayashi


  • Epilepsy in Children With Congenital Heart Disease: Risk Factors and Characteristic Presentations

    by Takashi Shibata, Maiko Kondo, Yosuke Fukushima, Mari Akiyama, Tomoyuki Akiyama, Teruko Morooka, Kenji Baba, Shinichi Ohtsuki, Hirokazu Tsukahara, Shingo Kasahara, et al.


  • Comprehensive study of metabolic changes induced by a ketogenic diet therapy using GC/MS- and LC/MS-based metabolomics

    by Mari Akiyama, Tomoyuki Akiyama, Daisuke Saigusa, Eiji Hishinuma, Naomi Matsukawa, Takashi Shibata, Hiroki Tsuchiya, Atsushi Mori, Yuji Fujii, Yukiko Mogami, et al.


  • Open Access

    Folic acid inhibits 5‐methyltetrahydrofolate transport across the blood–cerebrospinal fluid barrier: Clinical biochemical data from two cases

    by Tomoyuki Akiyama, Ichiro Kuki, Kiyohiro Kim, Naohiro Yamamoto, Yumi Yamada, Kazuya Igarashi, Tomohiko Ishihara, Yuya Hatano, Katsuhiro Kobayashi

    Objective: The use of folic acid (FA) has been discouraged in cerebral folate deficiency (CFD) because, theoretically, it could inhibit the transport of 5‐methyltetrahydrofolic acid (5MTHF) across the blood–cerebrospinal fluid (CSF) barrier. We present the clinical biochemical data of two cases with CFD to support this hypothesis.Methods: We measured CSF and serum 5MTHF concentrations in a patient with Kearns‐Sayre syndrome (KSS) and a patient homozygous for MTHFR C677T polymorphism before and during folate supplementation therapy. To evaluate these 5MTHF concentrations, we also analyzed CSF and serum samples in pediatric patients without folate supplementation.Results: Both patients had low CSF 5MTHF before treatment and high‐dose FA therapy did not normalize CSF 5MTHF. There was a dissociation between serum total folate and 5MTHF concentrations during FA therapy, which was considered to be due to the appearance of unmetabolized FA. The addition of folinic acid did not improve low CSF 5MTHF in the KSS patient and the cessation of FA resulted in the normalization of CSF 5MTHF. In the patient homozygous for MTHFR C677T, minimization of the FA dosage resulted in the normalization of CSF 5MTHF and an increased CSF‐to‐serum 5MTHF ratio.Conclusions: Our data suggest that excess supplementation of FA impaired 5MTHF transport across the blood–CSF barrier. In the treatment of CFD, supplementation of folinic acid or 5MTHF (in cases of impaired 5MTHF synthesis) is preferred over the use of FA. The reference values of CSF 5MTHF concentration based on 600 pediatric cases were also provided.