The movement problems caused by Parkinson’s can be improved by means of deep brain stimulation, but not every patient benefits from such an intervention. Surgery is even counterproductive for some patients. That is why it is important to ‘pick out’ those patients. The AI technology tested can improve that assessment.
Currently, the decision whether or not a Parkinson’s patient will benefit from brain surgery is made on the basis of a study that takes two and a half hours. During that examination, a neuropsychologist tests the entire thinking ability, such as memory, attention and language skills. However, the results of these tests are highly dependent on the patient’s education, culture and level of concentration
Assessing EEG with AI technology
“That is why there must be a quick, simple test that indicates when the patient is cognitively too bad for the operation,” said Victor Gereadts opposite Het Parool. With a brain film, or EEG, the electric impulses that nerve cells use to communicate with each other can be imaged in the form of moving waves. In Parkinson’s patients, whose thinking ability is impaired, those waves move more slowly.
In addition, an EEG can also measure which nerve cells all communicate with each other and how they respond to each other. However, these measurements cannot be visualized for assessment. That’s where AI technology comes into the picture. During the study, the algorithm used was able to distinguish 92 percent of patients with good and bad cognition.
The medical researcher used advanced AI technology from the automotive industry for his research. In new cars, this technology is used to detect and warn about hidden defects. “When you hit a post, sensors register the vibrations of the blow. A warning light in the car lights up if the invisible damage to the electronics is too great, so that you know it is better to drop by the garage, ”says Geraedts.
The vibrations that the sensors in a car detect are identical in terms of wave motion to the waves of an EEG. That is why this AI technology from the car industry appears to be useful in the assessment of EEGs in Parkinson’s patients.
Translate to EEG software
Geraedts has therefore demonstrated with his research that the AI algorithm works. Before this technology can be applied in practice, a lot of work still needs to be done. For the research and tests, Gereadts used a supercomputer that is not available for daily practice in hospitals. Therefore, the AI algorithm will first have to be built into the EEG software that doctors use in their doctor’s office.
“It now complements the existing assessment of cognition. We currently still assess the EEG waves manually, but hopefully soon the software will give an opinion on the patient’s cognition at the touch of a button, ”says Geraedts.
There are more developments in the field of using AI in the diagnosis and assessment of Parkinson’s. Earlier this year developed Maaike Dotinga, Technical Medicine student, developed an AI algorithm that was able to recognize Parkinson’s on images made with the SPECT-CT scanner during her graduation internship at Isala. With the help of this AI algorithm, a computer can now determine whether or not the patient has Parkinson’s with new images.