Differentiation of speech in Parkinson's disease and spinocerebellar degeneration using deep neural networks.

INTRODUCTION: Assessing dysarthria features in patients with neurodegenerative diseases helps diagnose underlying pathologies. Although deep neural network (DNN) techniques have been widely adopted in various audio processing tasks, few studies have tested whether DNNs can help differentiate neurodegenerative diseases using patients' speech data. This study evaluated whether a DNN model using a transformer architecture could differentiate patients with Parkinson's disease (PD) from patients with spinocerebellar degeneration (SCD) using speech data. METHODS: Speech data were obtained from 251 and 101 patients with PD and SCD, respectively, while they read a passage. We fine-tuned a pre-trained DNN model using log-mel spectrograms generated from speech data. The DNN model was trained to predict whether the input spectrogram was generated from patients with PD or SCD. We used fivefold cross-validation to evaluate the predictive performance using the area under the receiver operating characteristic curve (AUC) and accuracy, sensitivity, and specificity. RESULTS: Average ± standard deviation of the AUC, accuracy, sensitivity, and specificity of the trained model for the fivefold cross-validation were 0.93 ± 0.04, 0.87 ± 0.03, 0.83 ± 0.05, and 0.89 ± 0.05, respectively. CONCLUSION: The DNN model can differentiate speech data of patients with PD from that of patients with SCD with relatively high accuracy and AUC. The proposed method can be used as a non-invasive, easy-to-perform screening method to differentiate PD from SCD using patient speech and is expected to be applied to telemedicine.

Long-term effects of the gait treatment using a wearable cyborg hybrid assistive limb in a patient with spinal and bulbar muscular atrophy: a case report with 5 years of follow-up.

Background: Spinal and bulbar muscular atrophy (SBMA) is a progressive neuromuscular degenerative disease characterized by the degeneration of lower motor neurons in the spinal cord and brainstem and neurogenic atrophy of the skeletal muscle. Although the short-term effectiveness of gait treatment using a wearable cyborg hybrid assistive limb (HAL) has been demonstrated for the rehabilitation of patients with SBMA, the long-term effects of this treatment are unclear. Thus, this study aimed to investigate the long-term effects of the continued gait treatment with HAL in a patient with SBMA. Results: A 68-year-old man with SBMA had lower limb muscle weakness and atrophy, gait asymmetry, and decreased walking endurance. The patient performed nine courses of HAL gait treatment (as one course three times per week for 3 weeks, totaling nine times) for ~5 years. The patient performed HAL gait treatment to improve gait symmetry and endurance. A physical therapist adjusted HAL based on the gait analysis and physical function of the patient. Outcome measurements, such as 2-min walking distance (2MWD), 10-meter walking test (maximal walking speed, step length, cadence, and gait symmetry), muscle strength, Revised Amyotrophic Lateral Sclerosis Functional Assessment Scale (ALSFRS-R), and patient-reported outcomes, were evaluated immediately before and after gait treatment with HAL for each course. 2MWD improved from 94 m to 101.8 m, and the ALSFRS-R gait items remained unchanged (score 3) for approximately 5 years. The patient could maintain walking ability in terms of gait symmetry, walking endurance, and independence walking despite disease progression during HAL treatment. Conclusion: The long-term gait treatment with HAL in a patient with SBMA may contribute to the maintenance and improvement of the gait endurance and ability to perform activities of daily living. The cybernics treatment using HAL may enable patients to relearn correct gait movements. The gait analysis and physical function assessment by a physical therapist might be important to maximize the benefits of HAL treatment.