Two new artificial intelligence
models have been created by researchers at Örebro University that are capable
of accurately differentiating between patients with dementia, including
Alzheimer's disease, and healthy people by analyzing the electrical activity of
the brain.
According to Josérebro University
informatics researcher Muhammad Hanif, "early diagnosis is crucial in
order to be able to take proactive measures that slow down the progression of
the disease and improve the patient's quality of life."
In "An explainable and
efficient deep learning framework for EEG-based diagnosis of Alzheimer's
disease and frontotemporal dementia," scientists integrated temporal
convolutional networks and long short-term memory networks, two cutting-edge AI
techniques. By analyzing EEG data, the algorithm can virtually accurately
predict whether a person is ill or well. This research is published in
Frontiers in Medicine.
80% confidence in differentiating between healthy and sick
The approach reached over 80%
accuracy when comparing three groups: healthy, frontotemporal dementia, and
Alzheimer's. Additionally, the researchers employ an explanatory AI method that
displays the components of the EEG data that influence the diagnosis. This aids
physicians in understanding how the system arrives at its decisions.
In the second paper,
"Privacy–preserving dementia classification from EEG via hybrid–fusion
EEGNetv4 and federated learning," the researchers created a
resource-efficient, compact AI model that protects patient privacy and is less
than one megabyte in size. Frontiers in Computational Neuroscience published
this research.
Several healthcare providers can
work together to train the AI system without exchanging patient data thanks to
federated learning. The model gets over 97% accuracy despite the privacy
protection.
"Traditional machine learning
algorithms are hampered by privacy issues and frequently lack transparency.
Hanif, an associate senior lecturer in informatics at Örebro University,
states, "Our study aims to address both issues."
AI recognizes patterns in electrical signals from the brain.
AI recognizes patterns in electrical
signals from the brain.The researchers have been successful in integrating
various techniques for deciphering electrical impulses from the brain. The AI
can recognize patterns associated with dementia by splitting EEG signals into
different frequency bands, such as alpha, beta, and gamma waves.
The algorithms are able to identify small variations between diagnoses and
long-term changes in the data. Furthermore, explainable AI technology guarantees
that the system is no longer a "black box"—it makes its
decision-making process transparent.
In this research, the scientists
show how AI might develop into a quick, affordable, and private method for
early dementia diagnosis. EEG is already a low-cost, straightforward technique
that can be applied in primary care. This opens up the possibility of broader
application in healthcare, from specialized clinics to future home testing,
when combined with AI models that can operate on portable devices.
In the future, the AI test might be utilized at home.
"Implementing preventative
actions that delay the progression of disease and enhance quality of life
requires early diagnosis. According to Hanif, "if solutions like this are
fully implemented, it could ease the burden for everyone involved—patients,
care staff, relatives, and health care professionals."
Researchers from Örebro University and a number of foreign organizations,
including universities in the UK, Australia, Pakistan, and Saudi Arabia,
collaborated to perform the studies.
"We intend to carry out
additional study by investigating more EEG variables, extending to larger and
more varied datasets, and include other forms of dementia as Lewy body dementia
and vascular dementia. Hanif says, "We will use explainable AI and make
sure that patient data is strictly protected."
0 Comments