How AI Can Detect Low Sugar Level in Humans Without Blood Sample?
Artificial Intelligence (AI) in healthcare is strengthening its presence with new capabilities to diagnosis the health conditions of people with an acceptable level of accuracy. Recently we have shared post explains how does Google AI detect breast cancer better than radiologists.
And now, a new study came out in which researchers developed a new AI-based technique that can detect the low sugar levels from raw ECG signals using the wearable sensors without any fingerpicking test or taking the blood samples.
Currently, apart from giving the blood samples at diagnosis centers, diabetic people use to measure glucose using the needles at their home with repeated fingerpicks over the day. This is a painful process deterring patient compliance.
But now a new technique developed by researchers at the University of Warwick works with an 82% accuracy, could be the best option for invasive finger-prick testing with a needle, especially for kids who are afraid of needles fingerpicking.
How AI Detect Low Sugar Levels or Hypoglycemia?
The University of Warwick researchers examined how ECG readings changed during a hypoglycaemic event, when blood sugar levels fall below four millimoles per litre.
And they then used the AI system to recognize low levels compared to normal readings.
As per Dr Leandro Pecchia, from Warwick’s School of Engineering, the AI model has been trained to detect such hypoglycaemia via few ECG beats. This is possible as ECG can be detected in any circumstance, even while sleeping.
How AI is Trained to Detect Sugar Level in Humans?
The study is also explained how the AI was trained on the specific patterns of individual patients rather than on cohort data. This is because the ECG signals that correspond to hypoglycaemic events are different for each patient.
The figure shows the output of the algorithms over time — the green line represents normal glucose levels, while the red line represents the low glucose levels.
However, this makes difficult, or almost impossible to develop a single AI algorithm that could be rolled out for all patients. However, this may restrict the technology to some degree but allow more tailored treatment of individual patients based on their personal ECG data.
This result is possible because the Warwick AI model is trained with each subject’s own data. Intersubjective differences are so significant, that training the system using cohort data would not give the same results.
Likewise, personalized therapy based on our system could be more effective than current approaches.
Other AI Articles Trending:
How Does AI Work in Radiology: Applications and Use Cases
How Does AI Detect Cancer in Lung Skin Prostate Breast and Ovary?
However, more precise and magnitude of training data will help the AI model learn to detect with more variations while diagnosing such diseases.
Getting healthcare training data is crucial for AI engineers to develop such models that can detect the various types of diseases among humans.
To avoid diabetes and keep your body fit, make sure to do some workouts and regular exercise and control your sugar intake.
A high or low sugar levels both are dangerous for health and if not cured with precautions, it can become life-threatening diseases.
This story was originally posted on www.vsinghbisen.com