Five months ago, my cousin walked into our grandmother's hospital room. As he dropped his bags on the floor and moved to sit on the couch, our grandmother followed him with a blank stare. She hesitated for a moment, before asking tentatively, "I'm sorry. Who are you?"

Not being recognized by your loved ones is an extremely traumatic experience. Someone who has know you your whole life-- your mother, father, or caregiver-- is suddenly transformed into a stranger with only the vaguest inclination of who you are. As engineers and problem solvers, we wanted to make the transition and decline caused by Alzheimer's and dementia easier on patients and their families, which is why we created Pensieve.

How it works (aimed and implemented features)

Aim: The app has three main sections: photos, trivia, and quiz. In photos, the patient can browse through old family photos and recollect relationships that have already been seared into their long-term memory. In trivia, the user can go through historical facts and important family stories and memories (When did India gain independence? What was your first car?). And finally, in quiz, the user can test their memory of both these aspects (photos and textual trivia) in order to monitor memory decline.

After entering the app, the user is prompted to log in to Facebook. From there, they can search for specific friends-- family members, close acquaintances, or others-- and access and download this friend's photos to the app. The app then stores the photos along with the information of the person tagged in the photos. This is the process to add photos to the app. In addition, old family photos or non-digital photos can be photographed, scanned in and tagged as well. The tagging is used for categorization in the quiz section.

For trivia, we have chosen to interface the app with Twitter so that anyone close to the patient can add their own stories or memories. By tagging all relevant posts a specific, personalized hashtag, the user can use the app to search for the hashtag and add all tweets to a list of memories or trivia. These can be further tagged in the app and classified for use in the quiz.

Finally, for the quiz, the app takes information already stored by the user and turns it into engaging questions designed to jog the patient's memory. We hope to implement a machine learning algorithm that will provide the users with some of the same questions in different stages of the disease, so that doctors and family members can monitor the changes in memory functions over time, and have real-time, quantifiable data.

Share this project: