The incorporation of game elements into brain-computer interfaces (BCIs) training has shown promising results in enhancing subjects’ performance while using a BCI [1]. These positive effects are mostly observed in training protocols that last few sessions, lacking of longitudinal studies to evaluate their long-term effects.

The main objective of this study is to verify the feasibility of using motor imagery collaborative BCI coupled with a social video game to enhance the user learning. The training phase will span different weeks with numerous par- ticipants to exclude subject variability and interference.

We study the effect of population ageing on the development of patents relative to age-related diseases and assistive technologies targeted for the elderly. Patents relative to age-related diseases are identified leveraging both the International Patent Classification (IPC) and a keyword-based approach, while patents in assistive technologies are identified using the classification proposed by the World Intellectual Property Organization (WIPO). We construct a firm-level panel data and study whether the exposure to an older population pushes firms to develop more patents in these technological fields. To solve the problem of endogeneity and isolate the causal effect of population ageing on innovation targeted to the needs of the elderly, we use past fertility rates as instruments, relying on the idea that fertility rates impact population ageing while at the same time not being affected by innovation regarding agerelated diseases. We find that when a firm’s exposure to people 75 years old or older increases by one-standard deviation, the number of triadic patent families relative to age-related diseases increase by 6% of its standard deviation, the number of conventional assistive technologies increases by 16% of its standard deviation and the number of emerging assistive technologies increase by 34% of its standard deviation.

This work presents a novel multi-turn dialogue system designed to enhance the social and emotional support capabilities of Socially Assistive Robots (SARs) for older adults living independently. Implemented on the TIAGo++ robot and developed within a ROS environment, the system integrates user profiling—including personality, emotional state, and engagement—into interactions powered by the Llama3.2-1b Large Language Model. Key components include speech recognition, emotion and gaze detection, and adjustable robot personas for personalized dialogue. Preliminary evaluations highlight the system’s feasibility, with acceptable processing times and the potential for improved contextual interactions through user-aware adaptation.

The development of lower limb exoskeletons (LLE) has gained significant attention due to their potential applications in assistive technology and human augmentation. However, the application of these devices is currently limited to clinical and rehabilitation settings. To overcome this limitation, adaptive gait generation (AGG) is a critical component for enabling LLEs to operate safely and naturally in real-world environments. Current methods to AGG mostly rely on analytical solutions based on hand-crafted functions for describing the gait trajectory, which suffer from limited adaptability to complex environments and a manual tuning of the parameters. In this work, we propose an imitative approach to teach a lower limb exoskeleton how to walk from human demonstrations through improved Kernelized Movement Primitives (KMP). The gait is then adapted in real-time to the environment using an RGB-D camera mounted on the exoskeleton waist. We believe that the integration of exoskeletons with vision and artificial intelligence will foster the creation of a new generation of wearable robots capable of assisting older adults in their everyday life.