Instrumentation & Measurement Magazine 25-6 - 8

Fig. 5. Obtained results for the tremor test, based on [10]. (a) Trajectory
reconstructed in the time domain for all coordinates; (b) Frequency spectrum
for the x coordinate of the trajectory.
highlighting the robotic arm, the servo motor and the tag. The
anchors of the positioning system are not visible as they are
hidden inside the wooden structure. Fig. 4b shows a schematic
representation of the positioning system in which it is possible
to note the spatial arrangement of the anchors provided in [11].
To verify the reliability of the magnetic positioning system
in the diagnosis of PD, three tests about motor symptoms
were performed. In particular, a tremor test, a stop movement
test for rigidity and a finger tapping test for bradykinesia were
carried out.
The tremor test was initially performed by replicating what
happens in patients with action tremor. Action tremor occurs
when a voluntary movement is performed or when the patient
tries to maintain a position against gravity. To emulate the
symptom, the tag was moved through a known trajectory generated
with the robotic arm, and an oscillation at a frequency
of 10 Hz (characteristic frequency of the action tremor) was
superimposed on the trajectory by means of the servo motor.
The chosen frequency was derived from [8]. Fig. 5 shows the
obtained results. In detail, Fig. 5a shows the trajectory reconstructed
by the positioning system in the time domain for all
coordinates. Subsequently, Fig. 5b shows the frequency spectrum
for the x coordinate of the trajectory obtained by means
of a Discrete Fourier Transform. Only the x coordinate was analyzed
since the same considerations were applied to the other
coordinates. It is possible to notice how the 10 Hz component
is clearly visible in the spectrum; this indicates the reliability
of the positioning system in identifying the emulated tremor.
Subsequently, the stop movement test was used to analyze
the symptom of rigidity, a symptom that can occur in patients
during a movement. In detail, the trajectory imposed on the
tag was divided into two parts: initially a linear moving part
(with robotic arm) and a tremor superimposed at a frequency
of 2 Hz (with servo motor); subsequently, a locking part of
the movement with a tremor superimposed at a frequency of
5 Hz (with servo motor). The choice of the tremor frequencies
was derived in [12]. Fig. 6 shows the obtained results. Fig. 6a
8
Fig. 6. Obtained results for the stop movement test, based on [10]. (a)
Trajectory reconstructed in the time domain for all coordinates; (b) Spectrogram
for the x coordinate of the trajectory.
shows the trajectory reconstructed by the positioning system
in the time domain for all the coordinates. The imposed trajectory
provided for the movement of the tag mainly along the x
coordinate, weakly along the y coordinate while the z coordinate
is kept constant. So the only significant coordinates for
the test are the x and y coordinates. Subsequently, only for the
x coordinate, since similar considerations hold for the y coordinate,
Fig. 6b shows the spectrogram for the x coordinate of
the trajectory. The spectrogram clearly shows how only the
2 Hz component is present in the moving part (up to about 3 s),
while only the 5 Hz component is present in the locking part.
Also, in this test the reliability of the positioning system was
confirmed.
Finally, to evaluate bradykinesia, the finger tapping test
was performed. This test requires a patient to touch between
the thumb and index with the maximum amplitude and speed
possible. To emulate the test, two healthy subjects carried out
the experiment distinguishing a no-disease case from a disease
case. In detail, to carry out the no-disease case, the movements
(thumb-index touch) were performed regularly, and to emulate
the disease case, irregularities were introduced both in
terms of the amplitude and speed of the movements since such
irregularities are typical of patients with PD. By tracking the
tag (placed on the index of subjects), it was possible to extract
features for the two emulated cases (no-disease and disease)
and through artificial intelligence techniques to diagnose this
motor symptom.
A number of repeated tests equal to 17 and 20 were carried
out for the disease and no-disease cases, respectively.
For each test, the z component of the tag was considered as
the most significant for the analysis. Fig. 7a shows the average
spectra obtained by averaging over all the repeated tests.
It is possible to identify two areas of the spectrum in which nodisease
and disease cases can be discriminated. Specifically,
in the 1-4 Hz interval, the disease case has a greater spectrum
amplitude, while in 4-12 Hz the no-disease case is prevalent.
Consequently, the feature INT1-4
and INT4-12
IEEE Instrumentation & Measurement Magazine
were introduced
September 2022
Instrumentation & Measurement Magazine 25-6

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 25-6
