In our last blog post, we proposed two ways we could use personalized language-based exercises as a solution to neurodegenerative diseases. Here’s a quick summary of each of them.

  1. To facilitate early diagnosis of neurodegenerative disorders, we could administer regular assessments to individuals at potential risk for such disorders. These assessments would require patients to respond to images, and based on cognitive indicators from their responses, would be used to predict the risk of neurodegenerative disorders before their onset.
  2. To keep patients’ cognitive abilities sharp as a preventative measure against neurodegenerative disorders, we could administer regular linguistic exercises to individuals at risk for these disorders. These exercises would be fine-tuned based on the patients’ condition and would be adapted over time based on machine learning.

After the second stage of rapid evaluation, we looked over the feedback our peers gave us. We’ll highlight a few specific pieces of feedback that we received, as well as our response to this feedback.

Some students suggested that we narrow our focus down to a specific disease to ensure that we’re able to tackle the problem accurately. Having realized that it may in fact be difficult to tackle multiple diseases at once, we’ve decided to focus on Alzheimer’s. We’ve made this change in response to other feedback as well, which suggested that language tests might be insufficient to diagnose neurodegenerative diseases. Since research suggests that Alzheimer’s disease involves components of linguistic decline, we decided it was a good disease to focus on.

Other feedback questioned whether mental exercises can in fact guard against neurological decline. Fortunately, numerous studies have backed this effect. For example, a 2011 study in the Oxford Journal of Neurology examined the effects of “brain training” in curbing cognitive decline.

One suggestion that we received was to incorporate brain scans to see what changed throughout the progression of neurodegenerative disease. While this is certainly a potential way to gauge indicators of neurodegenerative diseases, we feel that it is beyond the scope of our project and may be too ambitious to tackle in conjunction with our other goals.

We also received several suggestions about incorporating more technology into our solution. One suggestion that we received was to carry out assessments digitally -- through web or mobile apps -- to make assessments easier to take and more time-efficient. We’ve decided to incorporate this idea into our solution, as we believe it will make our goal of giving regular assessments more attainable, as well as making data analysis more feasible. This would also open up the possibility for automatically adapting patients’ exercises or assessments as they complete more of them. A related suggestion was to use a database to allow for easier diagnosis in the future. This suggestion would also augment our capabilities for collecting and processing data in a manner that allows us to perform analytics, so we’ve decided to incorporate it into our solution as well. We also received feedback suggesting that administering regular assessments to many people might be costly and impractical. This has also contributed to our decision to conduct our assessments and exercises digitally.

Another piece of feedback we received, particularly in response to our first proposed solution, was that early detection of neurodegenerative disorders may be insufficient to actually move closer to a solution to these diseases. While we initially intended for our first solution to be more of a diagnostic tool to allow for early treatment, we’ve realized that there’s no reason we can’t tie our two solutions together: create a regular assessment for patients to take that is both diagnostic and preemptive. Therefore, we’ve decided to merge our two solutions. Given all of our feedback, we’re now looking to devise a method by which patients can regularly perform exercises electronically, where these exercises are then used to analyze Alzheimer's risk. We hope to make these exercises adaptive, having them change based on past patient performance on exercises and medical history.