P. Schoellauf, M. Mueller-Mezin, S. Poell, M. Muellner, S. Gagl, T. Bloder, M. Zusag, L. Wagner, Christoph Stepan
Cognitive and language impairments are common sequelae in patients who have experienced brain damage. Traditional rehabilitation methods often require intensive, long-term care, which may not be feasible for all patients due to geographical, financial, or time constraints. However, its effectiveness in real-world settings remains under-explored. Digital neurorehabilitation with myReha, facilitated by technology-based interventions, has the potential to improve accessibility and personalization of therapeutic strategies.
P. Schöllauf, M. Zusag, T. Bloder, L. Wagner, M. Cekolj, M. Müller-Mezin, A. Calleja-Dincer, C. Stepan
Individual therapy plans and high therapy intensity have been proven to be effective in the rehabilitation of cognitive deficits such as language. Computer-based therapy programs can achieve the necessary intensity and thus increase the effectiveness of therapy. Recent studies show the great benefit of digital and telerehabilitative interventions in patients with neurological disorders. Usage of myReha in an outpatient environment demonstrated efficacy for patients suffering from brain injuries, exhibiting significant enhancements across all evaluated therapeutic domains.
T. Bloder, P. Schöllauf, M. Zusag, L. Wagner, M. Müller-Mezin, C. Stepan
Cognitive impairments such as aphasia, attention, memory, or perception disorders are very common symptoms of neurological diseases after brain damage. Since patients need individual therapy plans and a high intensity of therapy in neurorehabilitation, the use of digital rehabilitation tools is seen as having great potential. This study evaluates the effectiveness of digital speech and cognitive therapy within a real-world mobile health data set using the digital neurorehab platform myReha. With this tablet app, patients receive customized exercise plans through artificial intelligence from a large catalog of over 35 language and cognition exercises with over 30,000 examples. These exercise plans can be used independently by patients both in the clinic and on an outpatient basis. This study includes real world data from 183 patients with cognitive deficits following brain injury more than four weeks ago who trained with the myReha app for 60 days. This study evaluated the efficacy of this form of therapy using the myReha app. The results demonstrated a significant enhancement in all evaluated speech and cognitive abilities over the intervention period.
M. Zusag, L. Wagner
We propose an automated pipeline for robustly identifying neurological disorders from interactive therapeutic exercises, which are gathered via the mobile therapy app myReha. The app captures speech and cognitive parameters from over 30.000 tasks in various scenarios. Users get immediate and highly accurate feedback for pronunciation and coherency for language tasks, while voice recordings are fed to a feature extraction pipeline in the backend. These features are then used to construct speech characteristics, which are highly indicative of different neurological disorders, such as acquired aphasia after stroke. The data is visually presented in a web application nyra.insights, which allows medical professionals to quickly derive recommendations for treatment and closely monitor outcomes. During the Show and Tell session, users can experiment with the interactive myReha app and experience the real-time speech analysis capabilities via the nyra.insights web platform.
L. Wagner, M. Zusag, T. Bloder
This paper presents a fully automated approach for identifying speech anomalies from voice recordings to aid in the assessment of speech impairments. By combining Connectionist Temporal Classification (CTC) and encoder-decoder-based automatic speech recognition models, we generate rich acoustic and clean transcripts. We then apply several natural language processing methods to extract features from these transcripts to produce prototypes of healthy speech. Basic distance measures from these prototypes serve as input features for standard machine learning classifiers, yielding human-level accuracy for the distinction between recordings of people with aphasia and a healthy control group. Furthermore, the most frequently occurring aphasia types can be distinguished with 90% accuracy. The pipeline is directly applicable to other diseases and languages, showing promise for robustly extracting diagnostic speech biomarkers.
M. Zusag, L. Wagner, P. Schöllauf, T. Bloder, M. Cekolj, M. Müller-Mezin, A. Calleja-Dincer, C. Stepan
Aphasia is a common and debilitating speech and language disorder affecting millions of people worldwide. The manifestation of aphasia can vary significantly from individual to individual, making it a challenge for healthcare professionals, particularly speech pathologists, to accurately diagnose and classify the disorder. While various efforts have been made to automate the detection and evaluation of aphasic speech, the task remains challenging. However, the use of advanced machine learning models for the detection of aphasia and other speech disorders has the potential to significantly reduce the demand on clinical resources and thus support clinical teams effectively.
