Alzheimer’s disease and related dementias (ADRD) negatively impact over 6 million Americans, with an annual societal cost of more than $1.3 trillion. Early diagnosis of ADRD allows clinicians to evaluate a patient's eligibility for both pharmacologic and non-pharmacologic treatments. Detecting ADRD in its early stages has been shown to slow its progression through these treatments, thereby extending the quality of life for individuals and their caregivers.. The earliest changes in behaviors include apathy, mild irritability, dysphoria, impaired insight socializing less, and exercising less. The human eye rarely detects these subtle changes, often misinterpreting them as normal aging.. Thus, ADRD often manifests 20-21 months after family members first recognize overt symptoms. 

Digital phenotyping involves collecting and measuring an individual's observable traits, behaviors, or characteristics in real time using data from personal digital devices like smartphones. This process has the potential to detect subtle changes that are often imperceptible through human interaction.

Our team has integrated digital phenotyping in a smartphone app, RealVision, to detect subtle behavior changes associated with pre-morbid ADRD, through a composite of:  (1) speech fluency, prosody, hesitations, and word errors, which are indicators of apathy (e.g., measured using accelerometer and GPS); (2) mild irritability (e.g., measured through facial cues and movement tracking); (3) dysphoria (e.g., measured using eye scan and facial cues), and (4) impaired insight (e.g., measured using microphone, accelerometer and eye tracking)

Karen Fortuna, Amrutha Nair, Miti Shah, Rebecca Liu, Hannah Crowe-Cumella