Smartphones Revolutionize Tracking Antipsychotic Side Effects

Antipsychotic Medications are used to treat serious mental health issues such as schizophrenia, bipolar disorder, and psychotic depression. While they are good at helping patients manage their conditions, these medications can also bring along some unwanted guests, namely movement disorders. These disorders include problems like shaky hands, restlessness, and unusual movements, collectively known as extrapyramidal symptoms (EPS). Research shows that about one in three patients taking antipsychotics may experience these side effects, which can make their condition even harder to bear.

#The Role of Anticholinergic Medications

To tackle these movement disorders, doctors often prescribe anticholinergic medications. These drugs can help reduce the shaking and other movement problems caused by antipsychotics. However, they come with their own set of issues. Patients may experience blurry vision, constipation, trouble with memory, and even a return of their original symptoms. In short, while anticholinergic medications may help with shaking, they can also cause new challenges for patients.

The best way to deal with EPS is to avoid using medications that lead to these side effects in the first place. If EPS shows up, it’s usually better to change treatments rather than continue with medication that is causing problems.

#Screening for Extrapyramidal Symptoms

There are various scales that doctors use to find and evaluate EPS. Some of these include the Simpson-Angus Scale, the Abnormal Involuntary Movement Scale, and others. Unfortunately, these assessments require a physical exam in person, which can be a hassle, especially during times like the COVID-19 pandemic when many doctors have turned to virtual appointments. This situation has made it clear that there is a real need for remote evaluation tools for EPS.

#The Need for Remote Evaluation Tools

The goal of recent research was to create a way to assess and predict EPS without needing an in-person visit. Researchers sought to use a smartphone or tablet’s built-in gyroscopic sensors to gather data and make predictions about hand tremors caused by EPS.

#Study Design

The research had two main parts. First, the team collected data from healthy individuals and patients with EPS to train their Computer Models. Then, they validated these models with data from a separate group of individuals. All participants in the study were adults aged 18 to 65, and they gave consent to take part in the research.

In the group without tremors, participants could not have any mental or neurological disorders. For those with tremors, the only requirement was that they were being treated with antipsychotic medications and showed signs of hand tremors during the study.

#Data Collection and Processing

Data was collected using a specific software on an iPad and recorded for acceleration, orientation, and angular velocity over a one-minute period. To keep things manageable, scientists split the collected data into smaller segments of 10 seconds each.

Once the data was collected, it went through various steps to prepare it for analysis. Each segment was processed, and information was saved for further evaluation. The aim was to create a computer model that could predict hand tremors based on the information gathered from these devices.

#Building the Computer Model

The model built for this research is called EDEPS, which stands for Early Detection of Extrapyramidal Symptoms. It uses machine learning algorithms to make predictions about hand tremors based on the data collected from mobile devices. Different types of algorithms were tested, but it turned out that a method called Random Forest was the most effective for both predicting overall tremor presence and measuring severity.

The model converts raw data into useful information through a series of steps, including transforming it into a format that is easier to analyze. By using data about the tremors, the model can help doctors conclude whether a patient has EPS and how severe it may be.

#Training and Validation of the Model

In training the model, the researchers used over 2,300 segments of data from both groups: those with tremors and those without. The team also carefully reviewed how well their models could predict outcomes to ensure accuracy. They combined data from both groups and adjusted the model whenever new data was introduced.

The researchers kept a close watch on how well their model was performing and made any necessary tweaks to improve its predictions. The goal was to make the model as accurate as possible in determining whether a hand tremor existed and how severe it was.

#Testing the Model

The effectiveness of the model was evaluated using a separate group of participants. They ran tests using a variety of approaches, including looking at all segments and just the first 10 seconds of data. The model was able to accurately assess hand tremors and even predict severity in many instances.

#Analyzing Power Spectrum Density

One fascinating finding from the research was related to the power spectrum density. The researchers discovered that patients with hand tremors had a noticeable peak in their data around 5 Hz. This means that there appears to be a specific frequency associated with hand tremors in patients taking antipsychotic medications.

#Comparison with Previous Studies

While there have been past efforts to measure tremors using different types of sensors or devices, this study stands out because it uses readily available technology like smartphones and tablets. Other studies have used wrist sensors or smartwatches, but these methods can be tricky to implement and require more effort from both patients and medical staff.

Some studies have pointed out that traditional methods of testing fine motor skills can be useful for screening. However, they may not provide precise quantitative results, making them less practical for ongoing assessment.

#Limitations and Future Directions

Like any study, this research has its limitations. The model was not tested against individuals who might have other types of tremors like essential tremor or tremors caused by withdrawal from alcohol. The hope is that this model will be able to distinguish between various types of hand tremors, but that remains to be seen.

The researchers also noted that the prediction accuracy could still be improved. While the model showed promise, there was a need for further refinements, especially when it came to specific scoring metrics.

In summary, this groundbreaking work could have significant implications for how we evaluate and treat hand tremors in patients taking antipsychotic medications. Who knew a smartphone could also help tackle mental health challenges? With continuing research in this area, we may find even better tools to help those dealing with EPS in the future.