The close link between cognitive decline and language has fostered long-standing collaboration between the NLP and medical communities in dementia research. To examine this, we reviewed over 200 papers applying NLP to dementia related efforts, drawing from medical, technological, and NLP-focused literature. We identify key research areas, including dementia detection, linguistic biomarker extraction, caregiver support, and patient assistance, showing that half of all papers focus solely on dementia detection using clinical data. However, many directions remain unexplored: artificially degraded language models, synthetic data, digital twins, and more. We highlight gaps and opportunities around trust, scientific rigor, applicability, and cross-community collaboration, and showcase the diverse datasets encountered throughout our review: recorded, written, structured, spontaneous, synthetic, clinical, social media based, and more. This review aims to inspire more creative approaches to dementia research within the medical and NLP communities.

When Whitney Bailey bought an Amazon Echo, she wanted to use the hands-free calling feature in case she fell and couldn't reach her phone. She hoped that it would offer her family some peace of mind and help make life a little easier. In some ways, she says, it does. But because she has cerebral palsy, her voice is strained when she talks, and she struggles to get Alexa to understand her. To make matters worse, having to repeat commands strains her voice even more.

Linguistic features have shown promising applications for detecting various cognitive impairments. To improve detection accuracies, increasing the amount of data or the number of linguistic features have been two applicable approaches. However, acquiring additional clinical data could be expensive, and hand-carving features are burdensome. In this paper, we take a third approach, putting forward Consensus Networks (CN), a framework to classify after reaching agreements between modalities. We divide the linguistic features into non-overlapping subsets according to their modalities, let neural networks learn low-dimensional representations that agree with each other. These representations are passed into a classifier network. All neural networks are optimized iteratively. In this paper, we also present two methods that empirically improve the performance of CN. We then present ablation studies to illustrate the effectiveness of modality division. To understand further what happens in Consensus Networks, we visualize the interpretation vectors during training procedures. They demonstrate symmetry in an aggregate manner. Overall, using all of the 413 linguistic features, our models significantly outperform traditional classifiers, which are used by the state-of-the-art papers.

Researchers at Toronto-based company WinterLight Labs have recently devised a machine learning method of predicting dementia that prioritizes particular variables when analyzing data, which could help to isolate the effects of potentially confounding factors. Alzheimer's disease and other types of dementia are a major worldwide challenge, leading to the death of one out of three seniors in the U.S. alone. While the causes of these diseases have not yet been fully grasped, they can have detrimental effects on speech, memory, orientation and other important cognitive abilities. WinterLight Labs is developing AI-based tools that could help detect and monitor Alzheimer's disease, aphasia, dementia, and other conditions that affect humans' cognitive abilities. The company has achieved very promising results, developing tools that can classify subtypes of aphasia with up to 100 percent accuracy and dementia with over 82 percent accuracy.

This is the second in a three-part series reporting from Toronto's booming Artificial Intelligence sector where new technologies are being pioneered that will permanently change all of our lives Just 45 seconds in the company of scientist Frank Rudzicz and his machines is all it takes to determine whether or not you are suffering from Alzheimer's disease. In that time, the complex Artificial Intelligence (AI) algorithms that the 37-year-old and his team have developed are able to pick apart your voice and predict the severity of the disease to an accuracy of around 82 per cent (and rising). First, there is your actual use of language. Alzheimer's sufferers tend to leave longer pauses between words, prefer pronouns to nouns (for example, saying "she" rather than a person's name) and give more simplistic descriptions, such as a "car" rather than the model or make. Then there is what Rudzicz calls the "jittter and shimmer" of speech; variations in frequency and amplitude.

Rudzicz, who is also an assistant professor in computer science at the University of Toronto, admits there are complex regulatory issues around the extent to which AI machines should be used to diagnose patients. Currently, his models are being piloted in the largest network of retirement homes in North America, and among elderly patients in Edinburgh and Nice, to collect data and train the machines to understand different languages and accents. At present, they are only being used only to map cognitive decline within existing patients rather than actually diagnosis new ones. "We have always been careful to position this as an assessment aid rather than straight diagnosis," Rudzicz says. "One of the main risks I see with AI in healthcare is people can put a lot of faith into it and discount other sources of evidence."

A retirement home in north Toronto is preparing to welcome an unusual resident: Ludwig, an artificially intelligent robot. Adorned with spiky mauve-coloured hair, green-tinged eyes and a few quirky facial expressions to mimic a range of emotions, the two-foot-tall robot is made to look and act like a little boy. By drawing his elderly neighbours into conversation, Ludwig's creators say he can track and monitor signs of Alzheimer's disease or dementia. He's so good, he can detect subtle changes in speech and vocal patterns that might escape retirement home staff, says Isaac Weinroth, executive director of One Kenton Place, where Ludwig will begin trials next month. "Even things like the time gap between verbs, or the use of verbs, or lack of verbs, the time gap between sentences, between words in sentences," he says.

Make room, stethoscope and otoscope. Artificial intelligence (AI) applications are increasingly among the physician's standard instruments,experts at the University of Toronto say. "With electronic records, you can use text algorithms to read a patient's history, review their genetic predispositions, and correlate the information to make predictions," says Dr. Frank Rudzicz. Rudicz is one of five experts exploring the issues of privacy, accuracy and accountability at The Robot Will See You Now – the Revolution in Artificial Intelligence and Medicine at U of T on April 5. A research scientist with the Toronto Rehab Institute and an assistant professor (status only) in the department of computer science at the University of Toronto, Rudzicz is also a project lead within a federally funded national research network in technology and aging known as AGE-WELL NCE.