ABSTRACTDevelopment of the cortical circuits for sensory-motor processing require the coordinated integration of both columnar and long-range synaptic connections. To understand how this occurs at the level of individual neurons we have explored the timeline over which vasoactive intestinal peptide (VIP)-expressing interneurons integrate into mouse somatosensory cortex. We find a distinction in emergent long-range anterior-motor and columnar glutamatergic inputs onto layer (L)2 and L3 VIP+ interneurons respectively. In parallel, VIP+ interneurons form efferent connections onto both pyramidal cells and interneurons in the immediate column in an inside-out manner. Cell-autonomous deletion of the fate-determinant transcription factor, Prox1, spares long-range anterior-motor inputs onto VIP+ interneurons, but leads to deficits in local connectivity. This imbalance in the somatosensory circuit results in altered spontaneous and sensory-evoked cortical activity in vivo. This identifies a critical role for VIP+ interneurons, and more broadly interneuron heterogeneity, in formative circuits of neocortex.

Mammalian neocortex is important for conscious processing of sensory information with balanced glutamatergic and GABAergic signaling fundamental to this function. Yet little is known about how this interaction arises despite increasing insight into early GABAergic interneuron (IN) circuits. To study this, we assessed the contribution of specific INs to the development of sensory processing in the mouse whisker barrel cortex, specifically the role of INs in early speed coding and sensory adaptation. In wild-type animals, both speed processing and adaptation were present as early as the layer 4 critical period of plasticity and showed refinement over the period leading to active whisking onset. To test the contribution of IN subtypes, we conditionally silenced action-potential-dependent GABA release in either somatostatin (SST) or vasoactive intestinal peptide (VIP) INs. These genetic manipulations influenced both spontaneous and sensory-evoked cortical activity in an age- and layer-dependent manner. Silencing SST + INs reduced early spontaneous activity and abolished facilitation in sensory adaptation observed in control pups. In contrast, VIP + IN silencing had an effect towards the onset of active whisking. Silencing either IN subtype had no effect on speed coding. Our results show that these IN subtypes contribute to early sensory processing over the first few postnatal weeks.

Theoretical models suggest that maintenance and updating are two functional states of working memory (WM), which are controlled by a gate between perceptual information and WM representations. Opening the gate enables updating WM with input, while closing it enables keeping the maintained information shielded from interference. However, it is still unclear when gate opening takes place, and what is the external signal that triggers it. A version of the AX-CPT paradigm was used to examine a recent proposal in the literature, suggesting that updating is triggered whenever the maintenance of the context is necessary for task performance (context-dependent tasks). In four experiments using this paradigm, we show that (1) a task-switching cost takes place in both context-dependent and context-independent trials; (2) task-switching is additive to the dependency effect, and (3) unlike switching cost, the dependency effect is not affected by preparation and, therefore, does not reflect context-updating. We suggest that WM updating is likely to be triggered by a simple mechanism that occurs in each trial of the task regardless of whether maintaining the context is needed or not. The implications for WM updating and its relationship to task-switching are discussed.

Latent learning occurs when associations are formed between stimuli in the absence of explicit reinforcement. Traditionally, latent learning in rodents has been associated with the creation internal models of space. However, increasing evidence points to roles of internal models also in non-spatial decision making. Whether the same brain structures and processes support the creation of spatially-anchored or non-spatial internal models via latent learning, is an open question. To address this question, we developed a novel operant box task that allows to test spatial and non-spatial versions of a flavour-based sensory preconditioning paradigm. We probed the role of the retrosplenial cortex, a brain area associated with spatial cognition and subjective value representation, in this task using precise, closed-loop optogenetic silencing during different task phases. We show that the retrosplenial cortex is necessary for both spatial and non-spatial latent learning in mice. We further demonstrate that the requirement of retrosplenial cortex is limited to the preconditioning phase of the task. Our results provide insight into the specific role of the retrosplenial cortex in latent learning, demonstrate that latent learning plays a general part in the creation of internal models, independent of spatial anchors, and provide a novel avenue for studying model-based decision making.

The superficial layers of the superior colliculus (SC) are highly visual and receive direct input from the retina. Nonetheless, neural activity in the superficial SC (sSC) is modulated by locomotion and pupil-linked arousal. Here we show that visual responses of neurons in the sSC are additionally modulated by reward delivered prior to the visual stimulus. We trained mice to perform a visual detection task and recorded the activity of neurons in the SC using two-photon calcium imaging and electrophysiological recordings using high-density silicone probes (Neuropixels). Neurons across all layers of the SC responded to various task events, including reward delivery. However, responses to events like licking or movements did not explain the visual response modulation by reward. Electrophysiological recordings showed that most of the reward modulation occurred in the superficial rather than the deeper layers of the SC. Neurons also exhibited modulation by pupil-linked arousal, which was independent of the reward modulation. Performance of a population decoder to detect visual stimuli improved significantly by reward modulation but not by pupil-linked arousal modulation. Our results indicate that behavioural factors other than locomotion and arousal modulate visual activity in the SC.