The Role of Gamification and Evolutionary Computation in the Provision of Self-Guided Speech Therapy

Introduction

This chapter outlines the development and evaluation of Ogma, an autonomous and intuitive mobile application that is capable of consistently generating speech and language therapy sessions for individuals with aphasia – deterioration in the ability to comprehend or convey language in a coherent manner. Aphasia is most commonly associated with and caused by cerebrovascular accidents (CVA’s), better known as ‘strokes’, though it can be caused by various injuries and illnesses that lead to some degree of brain damage – tumors, infections, neurological disorders, etc. Aphasia can affect an individual in terms of their spelling, reading, writing, articulation, recognition and expression of words or sentences.

Dysgraphia is an issue often associated with aphasic individuals, whereby they experience difficulties in writing, most often with handwriting, but also in terms of their coherence. Dysgraphic individuals can present with issues related to their visual/spatial awareness, fine motor skills and or comprehension. Dysgraphic individuals tend to omit or substitute letters, whilst their handwriting exhibits frequent stylistic inconsistencies and functional issues, such as a lack of speed and composure. Dysgraphic individuals’ writing may be aided via the use of note-taking software and applications or specific writing utensils and materials, such as wide-grip pencils/pens, graph paper, paper with large spacing, or paper with raised lines.

Ogma – named after the Celtic mythological character that was attributed with the creation of the 4^th century Ogham alphabet – consists of both a minimal front-end mobile application, featuring a word recollection and identification game, and an offline genetic algorithm (GA). The GA is responsible for deciphering and gauging the complexity of discrete words/terms, and creating sessions, which are tailored to the interests and capabilities of the target user. The sessions generated by the GA are presented sequentially to the user as a series of images, which they are then tasked with first of all identifying and then attempting to spell the target term. The application provides the user with an easy to access prompting sequence to aid their progress throughout each session. This prompting sequence attempts to mimic the behavior of an experienced speech and language therapist in so far as, when an individual is struggling in a clinical setting, a trained therapist is capable of eliciting responses from them via certain types of prompts. Ogma is designed to be used unassisted where possible, and so if a user is finding it difficult to either identify or spell a particular term correctly, they can invoke a number of prompts:

• 1.

  They are first provided with a contextual prompt, e.g. “The boy kicks the ____”. The application utilizes a text-to-speech engine and ‘speaks’ the word “blank” rather than the target term, i.e. “The boy kicks the blank”.

• 2.

  If the user is still unable to identify the target word, they are presented with the same contextual prompt, but in this case the word’s initial grapheme is shown and the user is also given a phonemic prompt, e.g. “The boy kicks the b___”.

• 3.

  Finally the user is provided with an anagram, where the target word’s letters are shuffled and displayed onscreen. A visual cue is presented alongside each letter, denoting whether the letter is in the correct position or not. This anagram is interactive, allowing the user to drag the letters into place, become familiar with the correct letter order/positions and then attempt to reproduce the target word.