Cognitive, Language, Learning Disabilities & Low Literacy

Behavioral data of individuals with autism spectrum disorder and healthy controls in photo-taking and image-viewing tasks are collected to build a classifier for automatic and quantitative screening of ASD. Participants were asked to take photos to reveal their attentional preference used to train a model. The model also incorporated temporal information in eye tracking data collected from participants viewing 700 images.

PDMove, a smartphone-based passive sensing system, compares diferent gait patterns before and after taking medication to facilitate medication adherence monitoring. PDMove passively collects gait data in a nonclinical daily-life environment using built-in inertial sensors of a smartphone, and afterward the gait preprocessor helps extract gait cycle containing the Parkinsonism-related biomarkers.

A dataset for passive sensing of body sounds is collected to train a privacy preserving neural networks for detecting Parkinson's disease. Voice samples with body and speech sounds is collected via a smartphone app. They first collect body sounds and then perform the sustained vowel sound of the letter A.

Recordings of therapeutic activities for assessment of austism spectrum disorder (ASD) are collected to introduce a computer-vision assisted interface for video based evaluations of social attention. Expert therapists performed professional therapeutic activities with children with ASD and typical development, composing of multiple tasks to see development of social interaction.

People with and without dyslexia played an online game while interaction data is collected which is analyzed to find patterns that could predict dyslexia. For every game played, right and wrong answer were collected, number of clicks, overall accuracy and other metrics were collected.

This work aims to create a machine learning readability model for English, developed specifically for the needs of adults with autism and evaluate it on the ASD corpus Participants were asked to read a small paragraph and were later asked 3 MCQs on it.

The study explores whether eye tracking techniques combined with machine learning can be used to develop screening models for identifying youngsters suffering from dyslexia. Data was recorded by presenting a set of 3 reading modules to each candidate and following their eye movements throughout the activity to detect learning disability.

For quantitative prosodic analysis of neutral, stress-modified and rhymed speech in patients with PD. Participant speech was recorded through microphone and sampled by audio interface M-AUDIO fast track pro.