2019 Conference on Implantable Auditory Prostheses

14-19 July 2019

Granlibakken, Lake Tahoe

Page 157

M66: CONTEXT AND USER REQUIREMENTS FOR DIGITAL SPEECH THERAPY FOR THE

SEVERELY HEARING IMPAIRED

–

FIRST RESULTS (THERESIAH)

Markus Meis, Joerg-Hendrik Bach, Andreas Becker, Kerstin Bilda, Alev Erem,

Tobias Feith, Stefan Goetze, Jonna Juers, Andreas Radeloff,

Benjamin Tschuschke, Laura Tuschen

HörTech gGmbH, Oldenburg, DEU: Evangelisches Krankenhaus, Oldenburg, DEU

University of Applied Sciences, Bochum, DEU: KIZMO GmbH, Oldenburg, DEU

Fraunhofer IDMT, Hearing, Speech and Audio Technology, Oldenburg, DEU

Universitätsklinik für Hals-Nasen-Ohren-Heilkunde, Oldenburg, DEU

Goals: The aim of THERESIAH is to develop a digital therapy and documentation system for pronunciation and

hearing training in at least severely hearing-impaired patients who suffer from speech and voice disorders due to

insufficient auditory input (audiogenic speech disorders).

Motivation and approach: The THERESIAH project is developing a digital system consisting of the following

modules: software modules for automated pronunciation evaluation using an automated speech recognition system

(ASR), elements of adaptive training programs with listening and pronunciation exercises (PC or Tablet

PC/Smartphone with headset), and the connection of electromyographic methods (EMG) for explorative

investigation of articulatory muscle functions. Overall, tools are to be developed that continuously objectify

articulation and pronunciation quality. Patients will have the opportunity to do exercises at home, self-determined

and with a high training frequency. In addition, it is planned to support phoniatrists and speech therapists in the

clinical environment and in postoperative outpatient therapy with new documentation functions.

Methods: N=12 medical experts (speech therapists, audiologists, and phoniatrists) were interviewed for the context

and user analyses. N=50 patients with severe hearing loss (PTA > 60 dB at frequencies 0.5, 1.0, 2.0 + 4 kHZ better

ear) and possible speech disorders were selected from a database of 1,700 hearing impaired patients. A total of

N=23 patients were interviewed. Qualitative in-depth interviews and interviews with (partially) standardized

questionnaires were conducted.

Context analyses: Data from disease-specific HrQoL questionnaires such as the Nijmegen Cochlear Implant

Questionnaire (NCIQ, scale 'Speech production'; Hinderink et al., 2000) and the German version of the Articulation

Handicap Index (AHI-12; Keilmann et al., 2017) showed that postlingually severe hearing impaired patients with a

progressive or acute pattern reported a comparatively low severity of symptoms in terms of speech production and

articulation. This was also supported by a study by Grundmann et al. (2019) with bimodal patients. Prelingually deaf

/ hearing impaired cochlear implant (CI) patients who were fitted with CI’s at a late stage were identified as the

primary target group. It was found that the THERESIAH system should be primarily used in outpatient speech

therapy and that clinical CI rehabilitation can initiate outpatient care. The trustworthy speech therapist-patient dyad

was assessed as essential and is to be supported by a digital therapy system, but not substituted.

According to expert interviews and literature reviews, the most common abnormalities are replacements or

omissions of phonemes, as well as backshifting of phonemes, an increased voice pitch, a disturbance of nasality,

an increased speaking speed, low sentence melody and other disturbances of prosody. Extensive diagnostic

measurements will be carried out with the planned patient group in order to quantify the suspected abnormalities

and, if necessary, to qualify further speech disorders.

User requirements analyses: Initial analyses of user requirements towards THERESIAH showed that a serious

gamification approach with a high-resolution display of learning progress was desired by the respondents and that

the incorrect or correct pronunciation of the patient should primarily be reported visually due to the model-oriented

intervention approach, based on the Speech Motor Control Model from Perkell et al. (2000). In addition, EMG

measurements were accepted as an additional diagnostic and training option in the clinical environment.

Measurements are currently being carried out with a THERESIAH prototype in order to iteratively further develop

the system. First results of this second iteration involving further prelingually deafened patients will be reported and

discussed with regard to further development.

This work will be supported by the Federal Ministry of Education and Research

–

BMBF, THERESIAH

(13GW0209A)