1. Interneuron diversity across neocortical areas
We used mouse genetics to generate quantitative maps of the distribution of distinct classes of interneurons across the entire cortex. We investigated the areal and laminar distribution of GABAergic interneurons throughout the entire mouse neocortex. We generated maps for all GABAergic neurons, for interneurons derived from the MGE/POA or the CGE/POH, and for five specific subtypes. We also estimated the density of excitatory pyramidal cells in each area and calculated the ratio of excitatory and inhibitory neurons. Our analysis of these datasets indicates that the density of different subtypes of interneurons varies extensively across different neocortical areas, and so is the ratio of excitatory and inhibitory cells.
We studied whether programmed cell death plays a critical role in establishing the heterogeneous patterns of interneuron distribution across the neocortex. We found that the patterns of cell death are very heterogeneous across cortical regions. However, preventing programmed cell death is not sufficient to homogenise the density of interneurons across cortical regions, which suggests that interneurons are already deployed in different cortical areas in a heterogeneous manner. We are currently preparing a manuscript describing the heterogeneous distributions of interneurons across the neocortex and the role of programmed cell death in shaping the organisation of inhibition in the developing neocortex (Paul et al., unpublished).
We also investigated the role of pyramidal cell activity in the survival of CGE/POH-derived interneurons. We have found that the survival of both CCK basket cells and neurogliaform neurons relies on glutamatergic neurotransmission, as we previously described for MGE/POA-derived interneurons. In contrast, we found that the survival of VIP bipolar cells during early postnatal development does not depend on glutamatergic neurotransmission but rather on neuromodulation by incoming serotonergic inputs. This finding revealed that long-range connectivity can also impact the survival of cortical cells through the regulation of their activity during an early window of postnatal development (Wong et al., 2022). In a related study, we found that the activity of cortical pyramidal cells also influences the final density of striatal interneurons, which also undergo extensive programmed cell death during early postnatal development (Sreenivasan et al., 2022).
2. Early cortical wiring
We established two-photon calcium imaging to investigate spontaneous and evoked activity patterns in the neocortex during postnatal development. In these experiments, we performed cranial windows to record chronically from the same genetically defined interneurons in head-fixed mice from P6 to P20. These experiments aimed to elucidate how preventing programmed cell death impacts network organisation and dynamics during ongoing activity, as well as its modulation by sensory stimulation and sensory deprivation, in both neocortical areas. We found that early-born SST+ interneurons play a prominent role in regulating the early patterns of synchronised activity, which is crucial for regulating interneurons' programmed cell death. In addition, we observed that SST+ interneurons regulate the maturation of PV+ interneurons, thereby controlling the transition from highly synchronous to sparse coding in the developing cortex (Mòdol et al., 2024).
3. Functional studies
The relevance of postnatal death of cortical neurons for the appropriate functioning of the cortex remains unclear. We generated animal models with abnormal ratios of cortical excitatory and inhibitory neurons and explored whether these alterations impaired sensory processing and learning. We abolished the programmed cell death of cortical pyramidal cells using Nex-Cre;Bak1-/-;Bax1fl/fl mice. Because this process is not homogeneous, preventing the programmed cell death of pyramidal cells has different consequences across cortical areas. To examine the functional consequences of this disruption, we established go/no-go sensory discrimination tasks with visual and texture (whisker) stimuli in head-fixed mice. We found that changes in the neuronal composition of the neocortex do not significantly affect perceptual learning and sensory discrimination. To examine how changes in cellular composition affect the transfer of learning in sequential training, we trained head-restrained control and conditional mutant mice on texture discrimination. Once animals had achieved good performance, we trained them in the visual discrimination task. Strikingly, we found that while control mice previously trained on the texture discrimination task learned the visual discrimination task significantly faster than naïve control mice, this transfer of learning effect was not observed in mutant mice. These results revealed that mice with too many neurons could not benefit from previous learning, which is consistent with neuronal alterations in high-order cortical areas such as the mPFC. We submitted a manuscript for publication in October 2023, and we are now working on revisions requested by the reviewers (Sreenivasan, Jamul et al., unpublished).
