Human–machine interaction (HMI) has been widely addressed in the literature and also encountered strong commercial interest in the last two decades: a number of advanced solutions have been proposed in a range of diverse application fields such as user/web interfaces, mobile computing and computer graphics and, from a wider perspective, robotics, ambient intelligence, entertainment and computer support to collaborative work/ learning. The basic HMI goal aims at improving the interactions between users and machines by making machines themselves more usable and receptive to the user’s needs. This essentially includes studies on models and theories of interaction, on methodologies and processes for designing interfaces, on developing suitable techniques for evaluating and comparing interfaces and, of course, on exploring new hardware devices and software frameworks. In particular, HMI researchers have been working hard to maximize the naturalness of the interaction itself by reducing the gap between the cognitive and emotional model behind human behavior and the machine awareness of ‘‘what is going on’’ during the task accomplishment. This asks for development of expert systems, able to manage large amount of information coming from sensory activity, to intelligently process it, and to promptly and knowledgeably respond to human actions according to natural interaction standards and by means of suitable actuary devices. Information processing therefore plays a central role from this perspective, operating at different levels, from multimodal digital data manipulation to semantic meta data processing, and necessarily encompassing the most challenging computational intelligence paradigms for contextual adaptation, social-emotional competence, and cognitive reasoning abilities. This Special Issue is fully devoted to propose the most recent and stimulating advances within the multidisciplinary area of cognitive and emotional information processing for HMI. It collects eight original contributions, which cover some of the aforementioned topics providing to the reader an insightful panoramic view of the research achievements and open issues in the field of interest. The present papers are the result of a rigorous review procedure applied to the fifteen articles initially submitted. At least three independent experts have been involved for each paper (more than 50 in total) and up to four review rounds have been performed before final acceptance for publication. The first two contributions deal with the processing of one of the widely used medium signals in HMI applications, namely speech, in order to suitably detect spoken activity in interactive environments. In the work of Principi et al., an innovative person activity detection algorithm operating in reverberated environments and in the presence of multiple sources is proposed. It is composed of two main stages, both operating in real-time: the speech enhancement front-end and the activity detector. The former is able to reduce the distortions introduced by room reverberation, thus improving the speech quality for each speaker and consists S. Squartini (&) Department of Information Engineering, Universita Politecnica delle Marche, Via Brecce Bianche 1, 60131 Ancona, Italy e-mail: s.squartini@univpm.it