Forty right-handed participants (20 men and 20 women; 21–40 years old) volunteered for the study. No participants had history of neurological or orthopedic disorders, and all had a normal or corrected-to-normal vision. They were classified as consistent right-handers according to the Edinburgh inventory  (all participants scored 90–100% on this scale). The purpose and procedure of this study were adequately explained to the participants, and written informed consent was obtained before enrolment. Our institutional ethical review board approved this study and performed procedures according to the Declaration of Helsinki.
The participants used a bp-prosthetic simulator (Fig. 1a). All participants used this simulator for the first time. This simulator consisted of a hand hook, socket, rod, cable, and harness. This simulator could be controlled by a cable attached to a figure-of-eight shoulder harness wrapped around the contralateral shoulder (Fig. 1b). The hand hook, which is connected to the cable, is voluntarily opened by pulling the cable and automatically closed by rubber band power, releasing the cable. All participants wore the simulator on their left upper limb. They wore it on their left upper limb because prosthetic control with the non-dominant hand was highly difficult and indicated new motor learning. When wearing the simulator, they grasped a rod by their left hand; then, their left forearm was attached to the socket, and the harness was attached to their right shoulder. Thus, the hand hook opened with left shoulder flexion and scapular abduction and closed with left shoulder extension and scapular adduction (Fig. 1c, d). A skilled occupational therapist adjusted the harness tension and hook direction in adapting the prosthesis.
Evaluation of functional performance
We used two functional tasks to examine the functional performance of the unilateral and bilateral manual dexterity during prosthesis use: box and block test (BBT) and bowknot task (BKT). The BBT was used in the evaluation of unilateral manual dexterity during prosthesis use and performed based on the previous report . Participants stood in front of a box divided into two square compartments on a stable table, one compartment containing a total of 150 wooden blocks (2.5 cm2). The height of the table was set to the same height as the BBT kit and participants’ navel. A nonslip mat was placed under the BBT kit. Then, participants transported as many blocks as possible from one compartment to another with only their left-hand hook within 60 s . The BBT score was the number of transferred blocks within 60 s.
The BKT was used in the evaluation of bilateral manual dexterity during prosthesis use. The BKT kit consisted of a box (20 cm × 10 cm × 15 cm) and five shoelaces of 15 cm length on each side. The height of the table was set to the same height as the BBT kit. A nonslip mat was placed under the BKT kit. Participants stood in front of a box placed on a stable table and knotted five shoelaces as fast as possible from far to near with the left-hand hook and normal right hand. The BKT score was the time from the start to completion of knotting the five shoelaces.
The BBT and BKT scores were measured twice in each test session, and the mean values were calculated as the score. Participants practiced for 5 min each before the start of the evaluation of the two tasks.
Evaluation of immersion
The Visual Analog Scale (VAS) was used in evaluating the immersion level during AO using VR and Tab. After AO, participants provided an immersion score on the immersive visual experience elicited by the AO. We obtained the immersion score using the VAS, which ranged from 0 (not completely immersive) to 100 (completely immersive). The immersion score was the average of the immersion during the video observation performed twice.
Video image and viewing
Participants observed the BBT and BKT videos of the prosthetic expert. We made four videos for AO training: two videos for viewing in the VR system and two videos for viewing on the Tab. Two videos for VR were recorded using a VR camera (Mirage Camera with Daydream, Lenovo). One of two videos for VR shows the first person performing the tasks of the BBT and BKT (Fig. 2 a, b); the other video shows the third person performing the tasks (Fig. 2 c, d). Two videos for the Tab were recorded using a camera installed in the Tab (Huawei d-tab d1-01H, Huawei).
The first- and third-person perspective viewing for VR was conducted using a head-mounted display (Mirage Solo with Daydream, Lenovo). The first- and third-person perspective viewing for the Tab was conducted using the Tab with 10.1-inch display (Huawei d-tab d1-01H, Huawei). When viewing the videos, we indicated that participants performed AO with motor imagery.
In this study, an intervention was divided into five groups: 1) VR1st, 2) VR3rd, 3) Tab1st, 4) Tab3rd, and 5) Control groups, consisting of 8 participants.
- The first group was the VR first-person perspective group (VR1st). VR1st used VR for intervention. The VR video was viewed using a head-mounted display. The VR video watched in 3D 180° vertically and horizontally. Participants watched the videos of prosthesis skills and imaged them as if they were performing prosthetic operations. The video was a point of view as if one was performing a prosthetic operation (first-person perspective) (Fig. 2a, b).
- The second group was the VR third-person perspective group (VR3rd). VR3rd used VR for the intervention. Participants watched videos of prosthesis skills and imaged them as if they were performing prosthetic operations. The video was a viewpoint from which movement of the scapula to the forearm was easy to observe when performing a prosthetic operation (third-person perspective) (Fig. 2c, d).
- The third group was the Tab first-person perspective group (Tab1st). Tab1st used the Tab for intervention. Participants observed the same videos as those in VR1st on a Tab and imaged them as if they were performing prosthetic operations. The video was a point of view as if one was performing a prosthetic operation (first-person perspective) (Fig. 2a, b).
- The fourth group was the Tab third-person perspective group (Tab3rd). Tab3rd used the Tab for intervention. Participants observed the same videos as VR3rd on the Tab and imaged them as if they were performing prosthetic operations. The video was a viewpoint from which movement of the scapula to the forearm was easy to observe when performing a prosthetic operation (third-person perspective) (Fig. 2c, d).
- The fifth group was the control group (Con). Con was evaluated for BBT and BKT without intervention.
Protocol (Fig. 3)
The participants were divided into five groups for the intervention. As a pre-intervention (Pre) evaluation, BBT and BKT were evaluated. The first and second interventions were Session 1 (S1) and Session 2 (S2), respectively. The groups, except for the control group, imaged BKT and BBT videos while watching them for 10 min. Participants were allowed to remove the simulator and move their hands during video observation. During video observation, they were instructed to imagine themselves as if they were performing a prosthetic operation.
We performed the BBT and BKT after the first and second interventions, which were Post 1 (P1) and Post 2 (P2), respectively. We obtained an immersive evaluation at P1 and P2 in four groups: VR1st, VR3rd, Tab1st, and Tab3rd.
To clarify the effects of intervention, we used the following: 1) BBT change rate, 2) BKT change rate, and 3) immersion score.
BBT change rate is an increased rate of the BBT scores of P1 and P2 when Pre is set to 0. BKT change rate is a decreased rate of the BKT score of P1 and P2 when Pre is set to 100. The immersion score was the average of the immersion during video observation of S1 and S2.
Initially, each data was confirmed for normality using the Shapiro-Wilk test. After normality was confirmed, BBT and BKT change rates were analyzed using two-way ANOVA for factors of groups (VR1st, VR3rd, Tab1st, Tab3rd, and Con) and sessions (Pre, P1, and P2). Immersion was analyzed using one-way ANOVA for factor of groups (VR1st, VR3rd, Tab1st, Tab3rd). If a significant main effect was found, the Holm’s method was used for post hoc comparisons. Then, we investigated the relationships among BBT and BKT change rates (Pre to P2) and immersion score using Spearman’s rank correlation test. All results are reported as means ± standard deviations. Statistical analysis was performed using IBM SPSS version 23.0. The significance level was set at 5%.