This algorithm could detect Parkinson’s during sleep

This algorithm could detect Parkinsons during sleep

  • News
  • Published on
    Updated


    Reading 1 min.

    in collaboration with

    Dr Gérald Kierzek (Medical Director of Doctissimo)

    Medical validation:
    August 31, 2022

    This artificial intelligence model could analyze breathing during the night and thus detect Parkinson’s disease early.

    In France, more than one million people are affected by a neurodegenerative disease. Among them, Parkinson’s disease – a degenerative pathology that results from the slow death of brain neurons – affects 25,000 new people each year. To detect it early, scientists compete in ingenuity, as evidenced by a recent study published in the journal Nature Communications.

    Detecting the disease during sleep

    The objective of this work? “Detect the disease and follow its progression from nocturnal respiratory signals”, note the scientists, thanks to an artificial intelligence (AI) model.

    Remember that Parkinson’s disease affects the way you breathe, long before other signs appear (tremors, muscle stiffness, slowness, fatigue, etc.), resulting in late diagnosis.

    This machine would therefore have many advantages: it would make it possible to reproduce the functioning of a human brain, while measuring the breathing patterns of patients while they sleep.

    This model can also estimate disease severity and progression “, add the researchers.

    In total, to arrive at these observations, 7671 patients – including 757 with Parkinson’s – tested this model of artificial intelligence.

    A non-invasive test

    Last advantage of this machine and not the least: it analyzes the breathing of patients in a non-invasive way – that is to say without any physical contact – unlike other current diagnostic techniques (MRI, analysis of fluid in the brain ).

    Our study demonstrates the feasibility of an objective, non-invasive, at-home evaluation of the model, and also provides initial evidence that this artificial intelligence model may be useful for risk assessment prior to clinical diagnosis. “, conclude the researchers.

    Finally, this model would make it possible to determine the progress of the disease. A great help for doctors.

    dts1