CogniSpeak: A Mobile Speech-Based Serious Game for Early Monitoring of Cognitive Changes in Older Adults and Individuals with Mild Cognitive Impairment

Abstract Background: Early detection and continuous monitoring of cognitive changes are critical for timely intervention in aging populations and individuals at risk of mild cognitive impairment (MCI). Speech-based digital biomarkers and serious games have emerged as promising, low-burden tools for cognitive assessment and engagement. Objective: This study introduces CogniSpeak, a mobile serious game designed to monitor and stimulate cognitive functions through natural speech-based interactions. The system aims to capture speech-derived cognitive markers during gameplay to support early identification of cognitive changes in older adults, including those with MCI. Methods: CogniSpeak is implemented as a mobile application (Android and iOS) featuring short, adaptive game sessions targeting memory, attention, executive function, and language. Players interact with the game using natural spoken Persian responses. Speech data are analyzed to extract linguistic and temporal features, including speech rate, pause patterns, lexical diversity, syntactic complexity, and semantic coherence. A pilot study design is proposed to evaluate feasibility, usability, and associations between speech-based features and in-game cognitive performance. Conclusions: CogniSpeak demonstrates the potential of speech-based serious games as scalable and non-invasive tools for early cognitive monitoring in aging populations. Methods This study proposes a pilot interventional design involving older adults from the general population as well as individuals with mild cognitive impairment. Participants will engage with the CogniSpeak mobile application over a period of four to six weeks. Each session lasts approximately 10–15 minutes and includes speech-based cognitive tasks embedded within a game environment. Speech data will be anonymized and processed to extract temporal, lexical, and syntactic features. In-game cognitive performance metrics and usability scores will be analyzed to explore correlations with speech-derived features.