The participants engage in cognitive exercises while cycling through two virtual scenarios: a park and an urban environment. Three cognitive tasks are available to perform while cycling. Each task has been designed to focus on one or more cognitive abilities and try to reproduce activities and situations of daily living as much as possible.

Each exercise is performed in couple, collaborating or competing with another peer. Participants communicate with each other thanks to a voice chat.

In the case of competitive gameplay, both participants are expected to perform the same task and respond to the same stimuli. Each participant can view their own score as well as the score of their opponent.

In the case of collaborative gameplay, the system allows participants to view the same environment and engage in the same task. However, participants take turns playing for a fixed number of seconds, determined by the therapist.

The same dual-task exercises described for the social modality can be played also in a single-player mode. Participants perform the cognitive task autonomosly and increase only their personal score.

Pairs of participants are formed, who will be the SocialBike players; these pairs consist of users from the same center. The two players are in the same room where two replicas of the SocialBike system are installed.

The participants can each view, on their own screen, the virtual scenario, either a park or a city environment. Movement within these VR environments is synchronized with the patient's actual pedaling speed, measured in revolutions per minute (RPM) from the cycle ergometer. Movement occurs along a predefined path. The participant's heart rate will always be monitored.

The therapist chooses the optimal ergometer workload, the type of exercise, and social/single-player mode. Each session (either single- or multi-player mode) lasts 10 minutes. The available exercise are: "Find the intruder" (go/no-go task); "Prepare your Meal" (spatial attention); "Watch the road" (working memory).

Situational Motivation Scale (SIMS; Guay et al. 2000). This questionnaire is composed of 16 items to be rated with a Likert scale going from 1 (corresponds not all) to 7 (corresponds exactly). Items are grouped in four different subscale (Intrinsic Motivation, Identified Regulation, External Regulation, Amotivation). Each subscale score can vary between 1 (minimum) and 7 (maximum).

System Usability Scale (Brooke, 1996) - The scale ranges from 0 to 100. Higher scores indicate better usability.

Ratings can be described as follows: "Excellent" (85-100); "Good" (70-84); "Average" (50-69); "Poor" (<50).