• evaluation of hand function
• fine dexterity
• gross dexterity
• trans - radial amputees
• validation
• virtual eggs test

The evaluation of hand function is of great importance to both clinical practice and research activities. Assessment tools are essential to provide the therapist or investigator with relevant and objective information concerning the patient's status, the effectiveness of the treatment program and the assistive technology prescribed. In relation to the hand function, the tests available in literature assess the manual dexterity of the patient, that is the ability to coordinate the movements of the hand and fingers to grasp and manipulate objects. There are two types of manual dexterity: gross and fine. Gross dexterity is involved in manipulating larger objects, and it includes the movement of the whole upper limb(s) with little precision. In contrast, fine dexterity is defined as precise motions associated with controlling small objects using the distal part of the fingers -usually thumb and index- and characterized by a high eye-hand coordination. There are several tests that measure manual dexterity. Most of them measure gross dexterity, for example the box and block or Minnesota Manual Dexterity (MMDT) tests, while few of them assess the fine dexterity. For this reason, it has been designed the Virtual Eggs Test (VET): it resembles the task of transporting fragile and robust objects, thus requiring both gross and fine dexterity. The Virtual Eggs are blocks (40x40x40mm, ~50 g) equipped with a magnetic fuse which exploits the attraction force between two magnets to maintain a fixed distance between two opposite walls of the block. When the grasping force exerted on the object is larger than the attraction force between the two magnets, the walls collapse and the object “breaks”. The breaking threshold (i.e. the fragility of the virtual egg) can be tuned by changing the strength of the magnets and/or their relative distance.
The test is composed by 11 Virtual Eggs (VE) with increasing breaking threshold s, a platform, a timer, and a software to analyse the data. The test is divided in two phases. In the first phase, the participant familiarises with all VE. In the second phase, the participant is asked to transport from one side to the platform to the other, for seven times, each VE without break it and as fast as possible. For each repetition, the participant presses the button to start the time, transports the VE and presses again the button to stop the time. The software registered the time for each transport. The test starts with the most robust VE (VE #11) and ends with the most fragile (VE #1). The examinator takes note of the breaking occurrences using the GUI of the software and supervises the execution. The outcome of the software is a .txt file, where there are the number of correct (no broken VEs) or incorrect (broken VEs) transport and the time of each transport. The correct transport outcome is then fitted using the Weibull psychometric function, that is characterized by two parameters α and Λ. α determines the slope of the curve and the deviation of the threshold value along the abscissa, while Λ is the higher asymptote. Concerning time of transport, for each VE we consider the minimum time among the seven repetitions (i.e. the fastest execution). The data are then fitted using a two lines function and the outcomes are the slope a and the mean b that are following defined. The data related to the robust VEs (from VE #11 to VE #6) are fitted using a horizontal line y=b, the constant term corresponds to the mean of minimum time of transport of the robust VEs. The data related to the fragile VEs (from VE #5 to VE #1) are fitted with line y=a(x-7.3) + b (where the starting point on the abscissa is the value of the last fragile VE and the value of y-axis is the mean value). The four parameters (α, Λ, a and b) are finally combined in two outcome measures: one related to the gross dexterity combining Λ and b, and one related to the fine dexterity α and a.
The aim of this thesis is to validate the test for trans-radial amputees wearing a myoelectric hand and improve its design based on the considerations shared by the therapists and clinicians during the experimentations. The VET is validated analysing the data of two different populations: the control population (i.e. able body) and the target (i.e. pathological) which includes the assessment with the MMDT, the self-report questionnaire ABILHAND and ranking in order to be able to compare the metrics in term of reliability and validity. The experimentation for the target population was performed at the INAIL Centro Protesi (Budrio, Bologna), involving 35 trans-radial, mono-lateral amputees, and users of myoelectric hand prosthesis. The data for the control population were acquired at the Biorobotics Institute of the Scuola Superiore Sant’Anna, involving 35 able body participants. For the control population, the test was repeated four times by each participant: two the day 1 and two at day 14.
The VET was validated in terms of: concurrent validity, construct validity, test-retest (reliability) and known group validity.
We found that Λ and b of the VET correlate with the outcomes of the MMDT (ρ = 0.12) and ABILHAND (ρ = 0.35). This means that the VET can assess the gross dexterity by means of the parameters Λ and b or their combination. In addition, we found that α and a of the VET correlate each other and with the ranking of performance obtained by observing the performance of the amputees in the manipulation of fragile VEs (ρ = 0.00). This result confirms that these parameters (or their combination) measure the fine dexterity.
The test-retest show that the parameters don’t change over the time for a given population (ρ >> 0.05), confirming the reliability of the VET. The known group validity show that the test is independent from the different levels of amputation (within the trans-radial amputation population) and from laterality of the amputation (dominant vs non dominant). Thus, the outcome measures of the VET only depend on the amputee’s improvement in the use of prosthesis. Finally, the target population proved to be significantly different from the control population (ρ = 0.00), thus the VET can discriminate between the two populations.
The second aim of my work was to improve the kit and the protocol. The main problem found by clinicians/therapists was to signal the breaking of fragile object. We implemented a red led, that light up when a magnet nearest a hall sensor, included in the PCB, during the breaking of walls. We also modified the design of Virtual Egg so that the VET complies with the CE directive for medical devices.
In conclusion, we validated and improved the VET for the evaluation of both gross and fine dexterity. Future works will be focused on the analysis of test for other impairments, such as post-stroke patients.