Final Report Summary - ADAM (The Adaptive Auditory Mind)
The goal of the ADAM project has been to investigate experimentally a novel view of hearing, where active hearing emerges from rapid interplay between adaptive sensory processes and goal-directed cognition. This view differs significantly from current approaches in both physiology and cognitive science. Physiology tends to emphasize feed-forward (bottom-up) feature extraction within a fixed network of cortical and sub-cortical areas, with well-defined properties that are fixed or evolve slowly over time. Cognitive theories emphasize that auditory behavior - and indeed all sensory processing - is shaped by higher (top-down) influences. These two facets are usually studied and conceptualized separately.
Our objective instead integrated the adaptive processes induced by cognition as an intrinsic part of sensory processing. This idea promoted a new understanding of how the active auditory system is able to cope with realistic auditory scenes, in situations where current artificial systems typically fail. A striking demonstration of the idea of an adaptive auditory mind advocated by the project has been a novel auditory illusion discovered during the project, described in a multi-disciplinary paper that combined psychophysics, brain imaging, and modeling (Chambers et al., Nature Communications 2017) and that can be experienced here: http://audition.ens.fr/dp/illusion(opens in new window). Another important objective of our project has been the establishment of a research program specifically on the adaptive processes responsible for intelligent and robust listening in difficult noisy and realistic acoustic scenes. There were two major components to this effort: (i) hiring creative scientists as faculty, PDOCs and Graduate students who rapidly engaged in fruitful interesting projects; and also the (ii) establishment of a fully functional platform for behavioral physiological research with ferret and potentially other species, in which the cognitive neuroscience projects could be conducted. The development of the platform facilities has been a substantial endeavor requiring planning and close interactions with a wide range of experts and staff at the Ecole Normale Superieure (ENS). Crucially, the future of the platform is now secure thanks to the commitment of both ENS and CNRS, who have hired one PI and one technician on permanent positions. Therefore, the ERC has uniquely allowed for the creation of a cutting-edge interdisciplinary platform which will benefit the host institutions for years to come. All the above objectives have been fully achieved, in addition to the deployment of new imaging technologies such as the functional Ultrasound.
We have also conducted research on numerous other important aspects of the project, specifically computational models, psychoacoustics, as well as contributing commentaries, reviews, and opinion articles on topics related to the core of the project. We have made substantial advances in each of the three Aims of the proposal: (1) to explore and exploit the perception of ambiguous stimuli to study the role of rapid adaptive plasticity in the formation and segregation of auditory sources. (2) To study the physiological bases of how acoustic perceptual sources emerge and become memorized. (3) And to combine all these ideas to formulate algorithms for the segregation of speech and music in realistic environments.
We have also established many collaborations in France, Europe, and indeed internationally. Some of these collaborations led to new projects including the European project on the Cognitive Control of Hearing Aids, as well as an ENS project on recording cortical activity using 2-photon imaging in mice. We have also collaborated on numerous computational projects in which we made available for detailed modeling animal recordings of responses to various stimuli including click trains and speech.
Our objective instead integrated the adaptive processes induced by cognition as an intrinsic part of sensory processing. This idea promoted a new understanding of how the active auditory system is able to cope with realistic auditory scenes, in situations where current artificial systems typically fail. A striking demonstration of the idea of an adaptive auditory mind advocated by the project has been a novel auditory illusion discovered during the project, described in a multi-disciplinary paper that combined psychophysics, brain imaging, and modeling (Chambers et al., Nature Communications 2017) and that can be experienced here: http://audition.ens.fr/dp/illusion(opens in new window). Another important objective of our project has been the establishment of a research program specifically on the adaptive processes responsible for intelligent and robust listening in difficult noisy and realistic acoustic scenes. There were two major components to this effort: (i) hiring creative scientists as faculty, PDOCs and Graduate students who rapidly engaged in fruitful interesting projects; and also the (ii) establishment of a fully functional platform for behavioral physiological research with ferret and potentially other species, in which the cognitive neuroscience projects could be conducted. The development of the platform facilities has been a substantial endeavor requiring planning and close interactions with a wide range of experts and staff at the Ecole Normale Superieure (ENS). Crucially, the future of the platform is now secure thanks to the commitment of both ENS and CNRS, who have hired one PI and one technician on permanent positions. Therefore, the ERC has uniquely allowed for the creation of a cutting-edge interdisciplinary platform which will benefit the host institutions for years to come. All the above objectives have been fully achieved, in addition to the deployment of new imaging technologies such as the functional Ultrasound.
We have also conducted research on numerous other important aspects of the project, specifically computational models, psychoacoustics, as well as contributing commentaries, reviews, and opinion articles on topics related to the core of the project. We have made substantial advances in each of the three Aims of the proposal: (1) to explore and exploit the perception of ambiguous stimuli to study the role of rapid adaptive plasticity in the formation and segregation of auditory sources. (2) To study the physiological bases of how acoustic perceptual sources emerge and become memorized. (3) And to combine all these ideas to formulate algorithms for the segregation of speech and music in realistic environments.
We have also established many collaborations in France, Europe, and indeed internationally. Some of these collaborations led to new projects including the European project on the Cognitive Control of Hearing Aids, as well as an ENS project on recording cortical activity using 2-photon imaging in mice. We have also collaborated on numerous computational projects in which we made available for detailed modeling animal recordings of responses to various stimuli including click trains and speech.