Periodic Report Summary 1 - PREDICTIVECODINGTIME (Interactions between predictive coding and predictive timing in audition: characterizing the role of rhythm in repetition suppression through entrained brain oscillations)
A fundamental principle of brain function is its ability to predict future events based on prior information. This concept may have important implications for the way we think about how information is coded in the brain, and important consequences for our understanding of some psychiatric diseases such as schizophrenia. However, a systematic investigation of the neuronal mechanisms subserving predictions is still lacking. Henceforth, the present project seeks to bridge between two important lines of research in current systems neuroscience in order to enlarge our understanding of the brain’s predictive representations of the environment. The first line investigates the ability of the brain to predict “what” will happen in the sensorium, which is based on memory trace formation through neuronal adaptation to stimulus statistics. The second line investigates the ability of the brain to predict “when” an event is likely to occur, which is based on low-frequency neuronal oscillations tuned to environmental rhythms. Both research lines are linked by a series of studies conducted both in humans and nonhuman primates, integrating information obtained at a global scale, observed with human electroencephalography, with that obtained at a local scale, observed with laminar profiles of post-synaptic potentials and neuronal spiking activity in macaques, aiming to characterize the role of brain oscillations entrained to stimulation rhythms in modulating neuronal adaptation to stimulus statistics along the auditory pathway.
So far, by the end of the outgoing phase of the project, we have been able to master the laminar electrophysiology technique, scarce expertise in the EU, and collect a complete dataset of recordings from the auditory cortex (core and belt areas) of a macaque that we trained to perform an audiovisual selective attention task. In the task, streams of tones and shapes were delivered simultaneously and the animal had to detect a target in the attended stream while ignoring distracting stimuli appearing in the unattended stream. Both streams consisted of trains of repeated stimuli presented in a rhythmic or non-rhythmic fashion. Preliminary analyses of the data show that, as expected, the behavior of the macaque replicates human behavior. That is, targets appearing after several repeated stimuli tend to be detected more often and faster than those appearing after just a few preceding stimuli and this correlation is more stable in rhythmic vs. non-rhythmic stimulus delivery modes, meaning that contextual rhythms organize the encoding of the content of stimulation. Importantly, preliminary analyses of the electrophysiological recordings in the auditory core (A1) show that the usual reduction of spiking with repeated stimulation is enhanced by rhythm as well, and that it is more prominent in the auditory attend condition for stimuli outside the best frequency of the tuning of the recording site and in extragranular layers of the cortex. This suggests that, while the feedforward input to A1 is reasonably preserved regardless of attention and environmental context (stimulus probability and timing), these factors do interact and shape neuronal activity acting as a spectrotemporal filter that sharpens response profiles. From a theoretical perspective, it implies that while a subset of neurons of the auditory cortex faithfully encode the incoming acoustic input, another reflects the likelihood and rhythm of the stimulation, and this can be interpreted as a sign that the brain has increased the confidence of its predictions of the environment. We are now analyzing the data to ascertain that the gain in repetition suppression in the rhythmic condition is due to the entrainment of neuronal oscillations shaping the excitability of neuronal ensembles coding for non-preferred stimuli in the extragranular layers of the cortex.
Besides keeping with the analyses of the non-human primate data, in the returning phase of the project we are going to perform an experiment with healthy human participants in which we will record their neuronal activity with electroencephalography (EEG) while executing exactly the same task that our macaque performed. This way we will be able to study the impact of global-scale cortical oscillations entrained to attended rhythmical stimulation on the encoding of probabilistic stimuli, which is indexed by several markers derived from analyses of the EEG (event-related potentials, gamma-band synchronization, etc. ), and complement it with the local scale we obtained with our animal. Thus, by the end of the project we hope we will have a broad picture of the neuronal mechanisms involved in forming expectations of the acoustic environment, which not only help in adapting our behavior efficiently but are essential to perception itself.
As there is an increasing number of EU research groups studying the predictive brain, but most of them have completely overlooked the role of timing predictability (i.e. rhythm), we hope that the final results of this project will contribute to develop of a new theoretical framework, enlarging our vision on how the brain perceives the world and adapts behavior in a predictive fashion. Importantly, our results may also have direct applicability for clinical research on several syndromes with high personal, social and economic impact, all shown to be related to dysfunctions in sensory-memory trace formation and rhythm encoding including: schizophrenia, dyslexia, specific language impairment, multiple sclerosis, epilepsy, autism, Parkinson’s disease, alterations in consciousness states and ageing, among others.
So far, by the end of the outgoing phase of the project, we have been able to master the laminar electrophysiology technique, scarce expertise in the EU, and collect a complete dataset of recordings from the auditory cortex (core and belt areas) of a macaque that we trained to perform an audiovisual selective attention task. In the task, streams of tones and shapes were delivered simultaneously and the animal had to detect a target in the attended stream while ignoring distracting stimuli appearing in the unattended stream. Both streams consisted of trains of repeated stimuli presented in a rhythmic or non-rhythmic fashion. Preliminary analyses of the data show that, as expected, the behavior of the macaque replicates human behavior. That is, targets appearing after several repeated stimuli tend to be detected more often and faster than those appearing after just a few preceding stimuli and this correlation is more stable in rhythmic vs. non-rhythmic stimulus delivery modes, meaning that contextual rhythms organize the encoding of the content of stimulation. Importantly, preliminary analyses of the electrophysiological recordings in the auditory core (A1) show that the usual reduction of spiking with repeated stimulation is enhanced by rhythm as well, and that it is more prominent in the auditory attend condition for stimuli outside the best frequency of the tuning of the recording site and in extragranular layers of the cortex. This suggests that, while the feedforward input to A1 is reasonably preserved regardless of attention and environmental context (stimulus probability and timing), these factors do interact and shape neuronal activity acting as a spectrotemporal filter that sharpens response profiles. From a theoretical perspective, it implies that while a subset of neurons of the auditory cortex faithfully encode the incoming acoustic input, another reflects the likelihood and rhythm of the stimulation, and this can be interpreted as a sign that the brain has increased the confidence of its predictions of the environment. We are now analyzing the data to ascertain that the gain in repetition suppression in the rhythmic condition is due to the entrainment of neuronal oscillations shaping the excitability of neuronal ensembles coding for non-preferred stimuli in the extragranular layers of the cortex.
Besides keeping with the analyses of the non-human primate data, in the returning phase of the project we are going to perform an experiment with healthy human participants in which we will record their neuronal activity with electroencephalography (EEG) while executing exactly the same task that our macaque performed. This way we will be able to study the impact of global-scale cortical oscillations entrained to attended rhythmical stimulation on the encoding of probabilistic stimuli, which is indexed by several markers derived from analyses of the EEG (event-related potentials, gamma-band synchronization, etc. ), and complement it with the local scale we obtained with our animal. Thus, by the end of the project we hope we will have a broad picture of the neuronal mechanisms involved in forming expectations of the acoustic environment, which not only help in adapting our behavior efficiently but are essential to perception itself.
As there is an increasing number of EU research groups studying the predictive brain, but most of them have completely overlooked the role of timing predictability (i.e. rhythm), we hope that the final results of this project will contribute to develop of a new theoretical framework, enlarging our vision on how the brain perceives the world and adapts behavior in a predictive fashion. Importantly, our results may also have direct applicability for clinical research on several syndromes with high personal, social and economic impact, all shown to be related to dysfunctions in sensory-memory trace formation and rhythm encoding including: schizophrenia, dyslexia, specific language impairment, multiple sclerosis, epilepsy, autism, Parkinson’s disease, alterations in consciousness states and ageing, among others.