The present study proposes a cognitive prosthesis device for face memory impairment as a proof-of-concept for the domain-specific cognitive prosthesis. Healthy subjects (n = 6) and a patient with poor face memory were enrolled. An acquaintance face recognition test with and without the use of cognitive prosthesis for face memory impairment, face recognition tests, quality of life, neuropsychological assessments, and machine learning performance of the cognitive prosthesis were followed-up throughout four weeks of real-world device use by the patient. The healthy subjects had an accuracy of 92.38 ± 4.41% and reaction time of 1.27 ± 0.12 s in the initial attempt of the acquaintance face recognition test, which changed to 80.48 ± 6.23% (p = 0.06) and 2.11 ± 0.20 s (p < 0.01) with prosthesis use. The patient had an accuracy of 74.29% and a reaction time of 6.65 s, which improved to 94.29% and 3.28 s with prosthesis use. After four weeks, the patient’s unassisted accuracy and reaction time improved to 100% and 1.23 s. Functional MRI study revealed activation of the left superior temporal lobe during face recognition task without prosthesis use and activation of the right precentral motor area with prosthesis use. The prosthesis could improve the patient’s performance by bypassing the brain area inefficient for facial recognition and employing the area more efficiently for the cognitive task.

Extra-spinal causes of sciatic pain are normally underdiagnosed, as they are extremely uncommon. Although pyriformis syndrome is recognized as one of the main causes of sciatic pain, other pelvic muscles that could cause sciatic pain are often overlooked. The present article describes a swollen inferior gemellus muscle with hematoma initially diagnosed with ultrasonography and later confirmed with magnetic resonance imaging (MRI) scan. Ultrasound revealed a swollen muscle with hematoma between the ischial tuberosity and the medial surface of the greater trochanter of the femur. MRI scan showed edematous change with an increased enhancement of the right inferior gemellus muscle. Ultrasound could be used to diagnose inferior gemellus pathology, but the muscle is easy to miss. Therefore, MRI could be preferred for conditions that impact deep or large areas in confirming inferior gemellus pathology.

Background: Dementia is a progressive neurodegenerative condition marked by cognitive decline and loss of functional independence. Among cognitive domains, executive dysfunction is a critical early contributor to reduced self-care capacity and increased caregiver burden. While cognitive assistive technologies have focused primarily on memory, few tools address executive function in real-time, daily tasks. To fill this gap, we developed a novel gamified cognitive prosthesis that integrates artificial intelligence (AI) and computer vision to guide users step-by-step through a simulated egg-cooking task. This system provides real-time audiovisual feedback to support planning, sequencing, and error correction. Objective: This study aimed to evaluate whether the AI-based cognitive prosthesis improves task completion time and executive function performance in individuals with mild dementia. Methods: We conducted a pilot study involving 12 patients with mild dementia and 7 age-matched healthy controls. Participants were asked to complete a 6-step gamified egg boiling task under 2 conditions: with and without guidance. The task was evaluated using a custom “Daily Task Completion Test” and a modified executive function performance test (EFPT) adapted to the cooking activity. Demographic and clinical data (age, sex, education, Mini-Mental State Examination, Clinical Dementia Rating, activities of daily living, instrumental activities of daily living, and Dementia Severity Rating Scale) were recorded. The System Usability Scale (SUS) was also collected postintervention. Results: In the mild dementia group, AI assistance significantly reduced median task completion time from 134.75 (IQR 92.50‐134.75) to 92.00 (IQR 65.00‐92.00; P=.03) seconds, and significantly improved the Executive Function Performance Test (EFPT) scores from 4.25 (IQR 1.75‐4.25) to 1.00 (IQR 0.00‐1.00; P=.005), reflecting a 31.7% improvement in efficiency and a 76.5% reduction in required assistance. No significant changes were observed in the control group. The mean SUS score was 80.53 (SD 24.97), indicating high usability. The AI system achieved a cumulative recognition precision of 0.93 (SD 0.07) and cumulative recall of 0.94 (SD 0.11). Conclusions: This pilot study provides preliminary evidence that an AI-based cognitive prosthesis can enhance executive function and task performance in individuals with mild dementia. The results support the feasibility of using real-time AI guidance in everyday tasks to promote independence. Given its modular design and promising usability profile, this system may serve as the foundation for future digital therapeutics targeting executive dysfunction. Larger, longitudinal studies are warranted to evaluate sustained cognitive and functional benefits.