Search

Background: Rhythmic Auditory Stimulation (RAS) is an effective technique to improve gait and reduce freezing episodes for Persons with Parkinson's Disease (PwPD). The BeatHealth system, which comprises a mobile application, gait sensors, and a website, exploits the potential of the RAS technique. This paper describes the tools used for co-designing and evaluating the system and discusses the results and conclusions., Methods: Personas, interviews, use cases, and ethnographic observations were used to define the functional requirements of the system. Low fidelity prototypes were created for iterative and incremental evaluation with end-users. Field trials were also performed with the final system. The process followed a user centered design methodology defined for this project with the aim of building a useful, usable, and easy-to-use system., Results: Functional requirements of the system were produced as a result of the initial exploration phase. Building upon these, mock-ups for the BeatHealth system were created. The mobile application was iterated twice, with the second version of it achieving a rating of 75 when assessed by participants through the System Usability Scale (SUS). After another iteration field trials were performed and the mobile application was rated with an average 78.6 using SUS. Participants rated two website mock-ups, one for health professionals and another for end-users, as good except from minor issues related to visual design (e.g. font size), which were resolved in the final version., Conclusion: The high ratings obtained in the evaluation of the BeatHealth system demonstrate the benefit of applying a user centered design methodology which involves stakeholders from the very beginning. Other important lessons were learned through the process of design and development of the system, such as the importance of motivational aspects, the techniques which work best, and the extra care that has to be taken when evaluating non-functional mock-ups with end users., Competing Interests: The authors have declared that no competing interests exist.

In this paper, a new supervised classification paradigm, called classifier subset selection for stacked generalization (CSS stacking), is presented to deal with speech emotion recognition. The new approach consists of an improvement of a bi-level multi-classifier system known as stacking generalization by means of an integration of an estimation of distribution algorithm (EDA) in the first layer to select the optimal subset from the standard base classifiers. The good performance of the proposed new paradigm was demonstrated over different configurations and datasets. First, several CSS stacking classifiers were constructed on the RekEmozio dataset, using some specific standard base classifiers and a total of 123 spectral, quality and prosodic features computed using in-house feature extraction algorithms. These initial CSS stacking classifiers were compared to other multi-classifier systems and the employed standard classifiers built on the same set of speech features. Then, new CSS stacking classifiers were built on RekEmozio using a different set of both acoustic parameters (extended version of the Geneva Minimalistic Acoustic Parameter Set (eGeMAPS)) and standard classifiers and employing the best meta-classifier of the initial experiments. The performance of these two CSS stacking classifiers was evaluated and compared. Finally, the new paradigm was tested on the well-known Berlin Emotional Speech database. We compared the performance of single, standard stacking and CSS stacking systems using the same parametrization of the second phase. All of the classifications were performed at the categorical level, including the six primary emotions plus the neutral one.