The Impact of Sensory-guided Behavior of Neuronal Processing in the Mouse Barrel Cortex and and a Novel Two-alternative Forced Choice Paradigm in Head-fixed Mice and Rats

Abstract

The reconfiguration of neuronal networks due to perceptual learning, and the impact of behavior in a task on online sensory processing are currently critical questions in neuroscience. In the work presented here, these topics were investigated in the barrel cortex of mice performing a sophisticated detection and discrimination task through the lens of two-photon calcium imaging. Layer 2/3 of the primary sensory cortex has been established as an essential locus of adult experience-dependent plasticity. Due to its unique topographical organization, the barrel cortex is an excellent area to investigate modulations of sensory processing in during whisker-dependent behavior. Before recording neuronal activity during learning of a sensory-guided task in project number two, we examined the ability of mice to perform a complex task successfully.
Thus, the primary goal of the first project was to establish if mice can learn complex behavior equally well as rats. In our two-alternative forced choice (2-AFC) paradigm, rodents learned to discriminate simultaneously applied single-whisker vibrotactile stimuli of different repetition rates. Both rats and mice could perform over 400 trials a day, with performance levels reaching over 90% of correct responses. The key finding of this project is that both species show comparable results regarding several behavioral readouts. Namely, their psychometric curves were similar with an average perceptual threshold of 53.0 Hz, and a 50.6- Hz frequency difference with corresponding Weber fractions of 0.58 and 0.56 respectively in rats and mice. Furthermore, their reaction times, the amount of omitted trials, rates of learning, and impulsivity were comparable among the two species. Additionally, we found that whisking before stimulus presentation impaired performance. A fundamental characteristic of our 2-AFC behavioral paradigm paired with whisker tracking is that it allows for a very precise control of the applied vibrotactile stimuli. Since the animal is head-fixed, additional recording and/or stimulating methods can be employed, such as electrophysiology or optogenetics. In the second project, its combination with long-term two-photon imaging will be presented. Furthermore, the task is very suitable for investigating interhemispheric interactions due to the possibility of bilateral stimulation. The main conclusion of the first project is that mice are equally adept as rats in learning complex behavior, such as the 2-AFC detection and discrimination tasks. This result opens up a multitude of possible research avenues by virtue of fusing behavioral measurements with potent recording and/or stimulating methods and taking advantage of the rich array of transgenic lines created in mice.
In the second project, we investigated plasticity effects due to behavior in a sensory-guided task, along with the impact of attentional engagement on sensory processing in the barrel cortex. To achieve this goal, our 2-AFC task was combined with long-term two-photon calcium imaging in layer 2/3 in the somatosensory cortex. This layer is a major site of adult cortical plasticity. However, our understanding of it is limited, since most studies focused on the granular and infragranular layers. Our data revealed strong plasticity effects due to perceptual whisker-dependent learning. During the baseline recording sessions, mice were passively exposed to sensory stimuli. Once the animals engaged in the operant conditioning sensory-guided task, the stimuli gained behavioral relevance by becoming associated with a water reward. Hence, layer 2/3 neurons display flexible modulation depending on the behavioral context, showing a more complex role for the primary sensory cortex than solely a passive relay of sensory information. Plasticity effects were equally reflected in comparing anesthesia recordings from before and after learning. A robust trend in sensory representation is the increase of evoked calcium transients with increasing repetition rates of the vibrotactile stimulus, also shown in this thesis. However, this monotonic increment was disrupted after the learning took place, which exclusively involved the highest and lowest frequencies. The middle frequency, never applied during the task, received fewer calcium transients than the lowest frequency, thus eliminating the distinctive property of vibrotactile frequency representation. Furthermore, we report the excitatory effect on the cortical representation of somatosensory stimuli during active behavior in the task, as opposed to passive stimulus exposure during unattended trials. Our bilateral 2-AFC paradigm allowed for the comparison of contralateral and ipsilateral responses, which was previously unknown in layer 2/3. We report that contralateral calcium responses are substantially larger than ipsilateral ones, contrary to the infragranular layers, where they are similar in size. To summarize the second project, behavior in a sensory-guided paradigm shapes both online sensory processing and long-term plasticity effects in layer 2/3 of the barrel cortex.