An intelligent hammer for remote examination of the tendon reflex

A whole new way of examining the tendon reflex is coming: an intelligent hammer capable of operating remotely. Typically the exam is divided into two main phases . Initially, the doctor strikes on the affected tendon area by stimulating a response from the patient and subsequently evaluating it with a certain score. This operation relies largely on the experience of the doctor, which is based on the feedback received from the "bounce" of the hammer on the affected part. Also, if the doctor does not hit the tendon correctly, the patient's response will be distorted and, therefore, not interpretable. For this reason it is quite complicated to perform the exam remotely.

An Artificial Intelligence instead of the doctor

Most of the so-called "non-urgent" medical examinations were postponed during the Covid-19 pandemic . This circumstance has made the need to develop more and more telemedicine tools even more urgent and evident in order to guarantee medical assistance to all those who need it. It is precisely from this need that a team of researchers from the Georgia Institute of Technology, led by prof. Jun Ueda, in collaboration with NITI-ON Co. and Tohoku University of Japan, has created a 'smart' hammer for the examination of the tendon reflex.

The smart hammer for examining the tendon reflex. Source: Georgia Tech Institute for Electronics and Nanotechnology
The smart hammer for examining the tendon reflex. Source: Georgia Tech Institute for Electronics and Nanotechnology

The intuition that prompted the researchers to create a system capable of performing the examination of the tendon reflex remotely is very simple. Since the physician must evaluate the patient's reflex type after the stroke, if this is not correct the reflex will not be reliable and the evaluation will be biased. If the hammer used, however, is equipped with a properly trained Artificial Intelligence (AI), he himself will inform the doctor of the goodness of the blow . The “brain” of this intelligent hammer is enclosed in a Machine Learning (ML) algorithm based on the Support-Vector Machine (SVM). It is adequately trained to be able to recognize a good blow from a wrong one simply by evaluating the acceleration of the impact, measured by an inertial sensor (IMU) placed on the head of the hammer itself.

Artificial Intelligence training

Scientists conducted trials on a group of 8 healthy adults (5 F, 3 M) with an average age of 36.5 years . Since one of the major critical issues of this test concerns the variability of the stimulation, two different types of experiments were carried out to train the algorithm using the hammer on a fake (latex) tendon to avoid recording the variability of the response movement. by the patient. In the first case, the acceleration values ​​of 50 shots were recorded correctly by expert and qualified personnel, in the second, instead, those obtained by non-expert personnel .

To measure the repeatability of the stimulation intensity, the scientists used the Relative Standard Deviation of the peak deceleration during impact. Finally, to model the component due to human error, the values ​​of a further 100 strokes were recorded, of which half on the tendon area (correct) and the other half on an adjacent area (wrong).

How the smart hammer works for the examination of the tendon reflex

Once the smart hammer hits the patient's tendon, an application installed on a smartphone or tablet sends the acceleration value via Bluetooth. The artificial intelligence then interprets the data and immediately informs the doctor whether it has touched the correct area or not . If so, a green LED also lights up on the hammer, which turns red if not. In addition, the doctor is invited to repeat the stroke several times until there are no more consecutive positive results, in order to limit the occurrence of false positives.

The intelligent hammer system. Source: A Smart Tendon Hammer System for Remote Neurological Examination
The intelligent hammer system. Source: A Smart Tendon Hammer System for Remote Neurological Examination

The results obtained are very promising. From the experiments conducted on the patient group, in 91.5% of cases the system was found to be reliable . In addition, the 'smart' hammer also allows beginners or ordinary people to learn faster to perform this exam correctly, thus creating a real learning aid. This allows anyone, especially family members, to guarantee medical assistance to patients suffering from neurological diseases, paving the way for telemedicine also to examinations of this type.

Article by Augusto Bozza

The articleA smart hammer for remote tendon reflex examination comes from Tech CuE | Close-up Engineering .