evaluations defined in UPDRS - Part III. Five main latent symptoms factors were identified and the correlations between the UPDRS scores assigned to the various tasks were reported. Similarly, in [104] a statistical analysis of the UPDRS motor scores was performed, using classical evaluation methods by neurologists and con-sidering PD patients in both ON (i.e., the intervals during which the medication is ef-fective) and OFF (i.e., the intervals during which the medication is not efef-fective) con-ditions, in order to identify latent relationships between UPDRS tasks and combine the tasks in “macro-groups” related to similar PD symptoms. Five of these groups, denoted as “factors,” have been identified (namely, gait/posture, tremor, rigidity, left extremities bradykinesia, and right extremities bradykinesia). The correlations (i) be-tween the UPDRS scores assigned by neurologists to the introduced “macro-groups”
of tasks and (ii) between them and an aggregate UPDRS score are also presented, showing that the macro-groups can be assessed separately and provide information about different aspects of the disease. In our opinion, a comparative analysis of multi-ple UPDRS tasks, based on a unified approach for both the kinematic characterization and the automatic UPDRS scoring, can be interesting in order to highlight their re-lationship with the different aspects of the PD and their contribution in the overall evaluation of the disease progression.
3.3 Experimental Set-up
3.3.1 Hardware
The BSN designed for the acquisition of the inertial signals used for the character-ization of the LA, S2S, and G tasks is formed by three Shimmer 2r nodes, one on the chest and one per thigh, attached to the body with Velcro straps and it is shown in Figure 3.1. We remark that the sensing platform, the devices’ positioning, the ac-quisition framework, and the setting configuration are the same used for analyzing singularly the G task, as explained in Section 2.3.
The placement of the sensors has been chosen taking into account two main mo-tivations: (i) the need to analyze the three tasks without changing the configuration of the nodes in order to minimize the patients’ stress and simplify the acquisition
Figure 3.1: The inertial BSN designed for the evaluation of the three UPDRS tasks of interest (LA, S2S, G): the subsets of nodes used in each task are marked with different colors.
procedure, allowing sequential execution of the three tasks; (ii) the higher accuracy and reliability of IMUs in measuring inclinations and accelerations, rather than posi-tions or displacements. With the current BSN configuration, indeed, all the kinematic parameters in the LA and S2S tasks are extracted from inclination and/or angular ve-locities measured with the nodes on the thighs and on the chest, respectively; at the opposite, in the G task, the majority of the features are extracted from acceleration signals directly measured by the sensor placed on the trunk.
3.3.2 Validation
As for the G task, the data acquired with the inertial BSN and the extracted kinematic features have been compared with those measured with the Vicon optoelectronic sys-tem, for validation purposes, also in the LA and S2S tasks. In [30], we have first demonstrated the equivalence between heel’ and thigh’ kinematics. More precisely, the three-dimensional orientations of the Shimmer nodes placed on the thighs are estimated, with reference to the Earth frame, through the Madgwick orientation
es-3.3. Experimental Set-up 79
timation filter (see Section 1.2.4). For each leg, the orientation component in the sagittal plane, corresponding to the inclination θ (dimension: [deg]) of the thigh, is extracted, together with the thigh’s angular velocity ω (dimension: [deg/s]). These signals are then compared to those estimated with the optoelectronic system, taking into account the angles and the angular velocities extracted by the three-dimensional position of reflective markers placed on the subject’s heels. The obtained results show a very high correlation between the two measurements (approximately equal to 0.98), motivating the use of θ and ω for the kinematic characterization of the LA task. The same approach has been applied, in the S2S task, for determining the correlation be-tween trunk inclinations estimated with the IMU and those extracted from the optical reference system, showing similar results [96]. The validation of the parameters used for the characterization of the G tasks has been discussed in Section 2.5.1. In general, the signals/parameters extracted through the IMUs and the Vicon system have a very high level of correspondence and can be considered accurate enough for the purposes of this work.
3.3.3 Subjects
For maintaining the consistency with the analysis performed for the G task, the set of subjects considered in this extended work has been kept unchanged and consists in 34 PD patients (22 males and 12 females). The average age is equal to 67.4 years (ages between 31 and 79 years) with a standard deviation equal to 11.6 years wheres the average Modified Hoehn and Yahr Scale score for the subjects was 1.6 (standard deviation equal to 0.47, minimum score score equal to 1, maximum score equal to 3) on the 1-to-5 scale (higher scores indicate more severe impairments and more advanced stages of the disease).
3.3.4 Acquisition Procedure
The positioning of the sensors on the patient’s bosy is shown in Figure 3.1. We tried to align the the x, y, and z axes of the Shimmer 2r node’s coordinate reference system, shown in Figure1.9 (a), to the upward-downward, right-left, and forward-backward
directions, respectively. In order to help clinicians to correctly align the nodes to the patient’s anatomical structure, the developed acquisition software has been provided with a functionality that allows to check the sensors’ placement: if the alignment is within a confidence range (heuristically defined), the examiner is allowed to proceed in the acquisition procedure; otherwise, a warning message is shown and the proce-dure is stopped until the sensors’ placement is correctly modified by the examiner.
The data acquisitions were carried out by asking the patient to execute the LA, S2S, and the G tasks sequentially. In each task, only the signals recorded by a proper subset of nodes of the BSN were considered—the subsets of Shimmer devices used for the evaluation of the single tasks are shown in Figure 3.1 using different colors.
For the LA task acquisitions, only the two devices placed on the thighs were used, whereas for the S2S task only the trunk-mounted node was considered. In the G task, all the BSN IMUs were used to achieve a complete characterization of the complex gait movement.
A total of 47 (94 for the LA task, considering separately RLA and LLA) acquisi-tions per task has been collected from the 34 patients, since some of them performed the sequence of tasks in distinct PD conditions (ON/OFF states) or at different times corresponding to different motor fluctuation phases. Under these conditions, indeed, the motor performance of a same subject and, consequently, its evaluation by the physicians, may vary in such a consistent way that considering each trial as an in-dependent sample does not introduce bias in the analysis. Although the number of patients has been kept the same considered in the previous chapter for the single evaluation of the G task, we remark that in the multi-task analysis the number of recorded trials has decreased from 55 to 47 for two reasons: (i) the 4 control patients included in the previous work have not been considered in the unified analysis; (ii) we kept only the trials in which the sequence of task, namely, LA, S2S, and G, was execute correctly.
The UPDRS evaluation of the collected trials has been carried out independently by three neurologist with expertise in movement disorders. A non-inter scale, with in-termediate scores (·.5), has been used–this allows the clinicians to increase the evalu-ation range and to label the trials in which they were undecided between consecutive