Biology subjects

Larkum, M. E.

Publications and source records attributed to Larkum, M. E..

18 recordsLinked to original sources

bioRxiv2026

Coordinated orienting behavior emerges following unilateral optogenetic activation of Superior Colliculus GABAergic neurons

Orienting coordinates movements of the eyes, whiskers, nose, and other effectors, but how the superior colliculus (SC) recruits these components over time remains unclear. We combined unilateral optogenetic activation of SC GABAergic neurons with high-speed tracking of facial movements and Neuropixels recordings to examine SC-evoked orienting in head-fixed mice. Stimulation recruited movements of the whiskers, nose, eyes, and pupil. Horizontal eye and nose movements were directed contralateral to the stimulated hemisphere, whereas vertical eye movements, pupil dilation, and whisker protraction did not differ significantly between stimulation sides. Nose and whisker movements generally preceded eye movements, followed by slower pupil dilation, revealing a temporally organized response. Recordings from the SC contralateral to the stimulated hemisphere revealed heterogeneous neuronal responses comprising short-latency excitation and delayed excitation or inhibition. Short-latency excitation was similar on trials with and without evoked movements, whereas delayed excitation was greater on trials that evoked eye or nose movements. Neurons modulated during spontaneous nose movements showed corresponding optogenetically evoked responses, linking stimulation-induced activity to endogenous movement-related activity. Thus, activation of SC GABAergic neurons initiates evolving population activity that recruits a coordinated sequence of orienting movements.

neuroscience↗

bioRxiv2026

Memory-based and surround-based pattern completion show distinct laminar profiles in human V1

Visual input is often incomplete, requiring the brain to infer missing information from prior experience or contextual cues. Establishing whether memory-derived and context-derived information shape pattern completion through shared or distinct top-down pathways is central to understanding how perception emerges from the integration of sensory input and prior knowledge. Across two experiments, we used 7T layer fMRI to measure activity in an occluded region of human V1 while participants viewed natural scenes. Memory-based completion was enabled by varying the level of exposure to full scene images before scanning, while surround-based completion was implemented via two scenes that were repeatedly presented in partially occluded form but never seen intact. In both fMRI experiments, category decoding in the occluded V1 region only emerged after multiple prior exposures, whereas novel scenes did not show reliable decoding. Repeated partially occluded scenes also supported decoding, despite the hidden region never being shown, consistent with surround-based pattern completion. Critically, mnemonic and surround-based pattern completion showed opposite laminar decoding profiles across V1 depths, with mnemonic feedback dominating deep layers, and surround-based feedback dominating superficial layers. Together, these findings suggest that completion in the early visual cortex depends on the sources of inferred information, revealing dissociable feedback pathways for memory- and context-driven inference.

neuroscience↗

bioRxiv2026

Behavioural Context Shapes Sensory Responses in Vibrissal Motor Cortex

Understanding how motor cortical circuits flexibly transform sensory and contextual information into behavior remains a central challenge. Whether neurons in primary vibrissal motor cortex (M1) multiplex across behaviors or are selectively engaged in context-specific actions is still unclear. To address this question, we trained mice on multiple vibrissal sensorimotor tasks, including a cue-triggered whisking-to-touch task and an air-puff-triggered licking task. Fast-spiking and regular-spiking neurons in layers 2/3 and 5 in vM1 responded robustly within [~]15 ms to air-puff stimulation. In contrast, these same neurons were only weakly modulated during goal-directed whisking-to-touch behavior. Unexpected air-puffs evoked responses in fewer neurons than expected stimuli. Trials in which stimulation elicited whisker movements produced smaller neural responses than trials without whisking. Stimulus-evoked activity in M1 was organized along a spectrum of response profiles with neurons exhibiting varying responses dynamics that cut across laminar and physiological distinctions. This organization of responses is consistent with context-dependent recruitment of M1 neurons. Together, these findings indicate that M1 activity is more closely associated with the selection of specific behavioral responses than with generalized sensory-motor encoding. SignificanceHow motor cortex links sensory input to behavior remains a central question in neuroscience. Do neurons that respond to vibrissal stimuli also participate in whisker-based behaviors, or do they reflect distinct functional states? Here, we show that activity in vM1 is context dependent. Across multiple behaviors, sensory inputs recruit different neuronal populations depending on behavioral context: air-puff stimuli evoke rapid and robust responses, whereas the same neurons are only weakly engaged during goal-directed whisking-to-touch. In addition, expected and unexpected stimuli activate partially distinct ensembles, and sensory responses are attenuated when stimuli directly trigger movement. These findings indicate that M1 does not uniformly encode sensory input; instead, activity reflects context-dependent action selection where neuronal populations are engaged according to behavioral demands.

neuroscience↗

bioRxiv2025

Shaft versus spine localization affects the structural plasticity of glutamatergic synapses

In the early stages of development, most excitatory synapses are formed directly on dendritic shafts. As neurons mature, these sites gradually shift from the shaft to dendritic spines. In fully developed excitatory neurons, the majority of glutamatergic postsynaptic sites containing the postsynaptic density (PSD) molecules reside on dendritic spines. However, some glutamatergic synapses remain as shaft synapses, yet their characteristics have remained unexplored. Here, we show that the molecular composition of the shaft PSDs closely resembles that of spine PSDs. Key components such as AMPARs, NMDARs, CaV1.2 channels, and F-actin interacting proteins, SynGAP, as well as cortactin, are present in comparable amounts in both synapse types. The major distinction between shaft and spine PSDs lies in the lower abundance of the scaffold proteins Shanks and Homer in shaft PSDs. Shaft synapses are not merely passive structures but actively participate in synaptic transmission. Their structure and function are modulated by changes in neuronal activity. Long-term live imaging combined with a cLTP protocol revealed that shaft PSDs were potentiated but rarely underwent a transition to spine synapses. In contrast, during LTD, shaft PSDs were eliminated more frequently than their spine counterparts. Together, these findings highlight excitatory shaft synapses as a distinct, and notably less stable, synapse type.

neuroscience↗

bioRxiv2025

mTORC1-Dependent Signaling in Layer 5b Neurons Is Required for Memory Consolidation

Memory consolidation relies on activity-dependent neuronal plasticity, particularly in deep-layer cortical neurons that integrate long-range inputs from medial temporal lobe (MTL) structures. In the neocortex, apical tuft dendrites of layer 5b pyramidal neurons receive direct input from the MTL within cortical layer 1, and their activation is critical for associative learning. However, the molecular mechanisms that enable learning-induced plasticity in these neurons remain poorly understood. Here, we demonstrate that activity-dependent, mTORC1-regulated translation in layer 5b neurons contributes to learning and memory. Using in vivo paradigms such as exposure to an enriched environment (EE) and in vitro models of network activation by chemical long-term potentiation (cLTP), combined with cell type-specific translational profiling (RiboTag), we identify candidate mRNAs associated with experience-dependent plasticity. We further show that pharmacological inhibition of canonical mTORC1 signaling during memory-relevant tasks abrogated memory consolidation. Our findings identify the mTORC1 pathway as a central regulator of learning-induced protein translation and cortical plasticity.

neuroscience↗

bioRxiv2025

Contextual modulation and blunted defensive responses to predators in head-fixed and freely moving mice

Behavioral responses to threat -- such as fleeing, freezing, or fighting--can be innate, learned, and strongly shaped by context or competing goals. Here, we asked whether exposure to an ecologically relevant predator obligatorily elicits canonical defensive behaviors across behavioral contexts. We examined predator responses in mice across four experimental conditions: one novel head-fixed reward-driven foraging task and three established paradigms in freely moving animals. In the head-fixed condition, water-deprived mice were trained to walk on a treadmill controlling a virtual environment and water reward delivery and were subsequently exposed to a live rat positioned above the lick spout. Despite the presence of the predator, most mice (5 of 7) maintained foraging performance at baseline levels. However, individual mice exhibited significant, coordinated changes in running speed, pupil diameter, eye movements, and posture, indicating engagement with the threat. To assess how context influences predator responses, we exposed 36 naive, freely moving mice to fear-inducing stimuli, including looming visual cues, rat odor, and a live rat. Even under these conditions, defensive behaviors were variable: only a subset of mice displayed avoidance or escape, and when presented with a freely moving rat, approximately half of the mice avoided the predator. Together, these findings show that predator threat does not elicit a uniform or obligatory defensive repertoire in mice. Instead, defensive responses are expressed flexibly, and are shaped by environmental constraints, task demands, and individual variability. These results challenge the assumption that innate fear behaviors are automatically triggered by predator encounters. Graphical AbstractUsing a head-fixed, reward-based foraging task and complementary freely moving paradigms, we show that mouse responses to predator-related stimuli are variable and context dependent, with limited expression of canonical defensive behaviors such as freezing or flight. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=200 SRC="FIGDIR/small/658679v3_ufig1.gif" ALT="Figure 1"> View larger version (65K): org.highwire.dtl.DTLVardef@a05a31org.highwire.dtl.DTLVardef@1815084org.highwire.dtl.DTLVardef@cd3a11org.highwire.dtl.DTLVardef@19f5d95_HPS_FORMAT_FIGEXP M_FIG C_FIG

animal behavior and cognition↗

bioRxiv2025

Tuft dendrites in frontal motor cortex enable flexible learning

Flexible learning relies on integrating sensory and contextual information to adjust behavioral output in different environments. The anterolateral motor cortex (ALM) is a frontal area critical for action selection in rodents. Here we show that inputs critical to decision-making converge on the apical tuft dendrites of L5b pyramidal neurons in ALM. We therefore investigated the role of these dendrites in a rule-switching paradigm. Activation of dendrite-inhibiting layer 1 interneurons impaired relearning, without affecting previously learned behavior. Remarkably, this inhibition profoundly suppressed calcium activity selectively in dendritic shafts but not spines while reducing burst firing. Moreover, excitatory synaptic inputs to tuft dendrites exhibited rule-dependent clustering. We conclude that active dendritic integration is a key computational component of flexible learning.

neuroscience↗

bioRxiv2025

Distinct roles of cortical layer 5 subtypes in associative learning

Adaptive behavior is critically dependent on associative learning, where environmental cues are linked with subsequent positive or negative outcomes. In mammals, primary neocortical sensory areas serve as pivotal nodes in this process, processing stimuli and distributing information to cortical and subcortical networks. Layer 5 (L5) of the cortex comprises two types of pyramidal projection neurons--intratelencephalic (IT) and extratelencephalic (ET) neurons--each with distinct downstream targets. Despite the crucial function of L5 as a main output node of the cortex, the specific contributions of these L5 neuronal subtypes to associative learning remain poorly understood. In the present study, by leveraging transgenic mouse lines, we distinguished IT and ET neurons in the primary somatosensory cortex and examined their roles in a whisker-based frequency-discrimination learning task. Longitudinal two-photon calcium imaging revealed distinct response characteristics between IT and ET neurons throughout learning. Interestingly, the activity of IT neurons hardly changed over the five days of learning, while the activity of ET neurons developed robustly. Furthermore, IT neurons appeared to show stimuli encoding from the beginning, whereas the ET neurons became increasingly responsive to stimuli associated with reward. Chemogenetic silencing of either IT or ET neurons both impaired learning, but in strikingly distinct ways, each associated with a different phase of learning. By modeling the response characteristics of IT and ET neurons using a reinforcement learning framework, we show that IT neurons primarily encode sensory stimuli, and their representations are critical for forming stimulus-reward associations. ET neurons instead represent the value of the stimulus, used for refining behavior. Thus, our results delineate the distinct roles of L5 IT and ET neurons, underscoring their integral and complementary contributions to associative learning.

neuroscience↗

bioRxiv2024

Bimodal dendritic processing in basket cells drives distinct memory-related oscillations

Neuronal oscillations are crucial for organizing distinct memory stages and behavioral states, yet the precise cellular mechanisms through which interneurons shape these dynamics remain incompletely understood. We propose and computationally explore a novel link between the nonlinear dendritic integration modes of parvalbumin-positive fast-spiking basket cells (PV+ FSBCs) and the modulation of hippocampal oscillations. Employing biophysical circuit level modeling, we test the hypothesis that PV+ FSBCs can flexibly adapt their firing patterns and influence rhythms, independent of gross changes in synaptic input quantity, by selectively engaging either their supralinear or sublinear dendritic branches. Specifically, supralinear dendrites promote high-frequency oscillations and reduce the excitation/inhibition (E/I) balance in the circuit, whereas sublinear dendrites enhance slow oscillatory power and elevate the E/I balance. This bimodal dendritic strategy provides PV+ FSBCs with an energy efficient mechanism to regulate circuit and oscillatory dynamics without necessitating large scale increases in synaptic drive. Our findings thereby uncover a previously unrecognized role for PV+ FSBC dendritic computations in regulating memory-related oscillations, offering new, experimentally testable hypotheses into the subcellular mechanisms that govern hippocampal rhythm generation.

neuroscience↗

bioRxiv2024

High-speed three-dimensional random access scanning with SPARCLS

High-speed volumetric imaging is crucial for observing fast and distributed processes such as neuronal activity. Multiphoton microscopy helps to mitigate scattering effects inside tissue, but the standard raster scanning approach limits achievable volume rates. Random-access scanning can lead to a considerable speed-up by sampling only pre-selected locations, but existing techniques based on acousto-optic deflectors are still limited to a point rate of up to [~]50 kHz. This limits the number of parallel targets at the high acquisition rates necessary, for example, in voltage imaging or imaging of fast synaptic events. Here we introduce SPARCLS, a method for 3D random-access scanning at up to 340 kHz point rate using a single 1D phase modulator. We show the potential of this method by imaging synaptic events with fluorescent glutamate sensors in mammalian organotypic slices as well as in zebrafish larvae.

neuroscience↗

bioRxiv2024

Experience-dependent predictions of feedforward and contextual information in mouse visual cortex

To navigate complex environments, the visual system must prioritize unexpected inputs or omissions over predictable sensory backgrounds. This saliency assignment is a hallmark of visual processing and depends on comparing visual feedforward inputs with learned contextual surrounds, a process often conceptualized as a difference computation. However, cortical circuitry poses fundamental challenges: visual contextual feedback is predominantly amplifying, and feedback-driven inhibitory neurons necessary for subtractive computations are relatively sparse. How then can complex contextual feedback be selectively compared with feedforward input at the level of individual neurons? Here, we show that divergent plasticity rules within single pyramidal neurons provide a solution to this problem. Using two-photon calcium imaging, we recorded layer 2/3 pyramidal cell responses in mouse primary visual cortex before and after visual familiarization. We dissociated feedforward and contextual signals by comparing responses to stimuli with and without an occluded sector. Repeated feedforward responses selectively weakened through adaptation and surround suppression, whereas contextual responses strengthened and generalized across scenes. These strengthened contextual inputs amplified non-adapted responses to novel images and drove neuronal activity when occlusion of feedforward input prevented surround suppression. This produced a conserved structure of feedforward and contextual responses across mice, monkeys, and humans. Ultimately, these results reveal a cellular mechanism for saliency assignment that does not require feedback-driven inhibition. Pyramidal neurons autonomously learn the generalized statistical structure of familiar contexts, allowing them to selectively amplify improbable inputs or omissions.

neuroscience↗

bioRxiv2023

Binge Feeding Promotes Appetite via Modulating Olfactory Flavor Representation

Binge eating commonly leads to overeating, but the exact mechanism is unclear. While it is known that experiencing flavor contributes to satiety, the interactions between flavor, feeding rate, and food intake remain unknown. Here, we demonstrate a novel feeding rate-dependent feedback loop between olfactory flavor representation in the anterior olfactory (piriform) cortex (aPC) and food intake. Using miniscopes for in vivo calcium imaging in freely foraging mice, we identified specific excitatory neuronal responses to food and water during slow feeding. Switching to binge feeding transformed these specific responses into unspecific global suppression of neuronal activity. Food consumption was predicted by the degree of suppression of neuronal activity in the aPC during binge feeding. Also, food deprivation enhanced neuronal activity suppression. We confirmed the hypothesis that aPC suppression promotes food intake with closed-loop optogenetics experiments. Together, we show that olfactory sensory representation in the aPC reciprocally interacts with consummatory behavior to enhance food intake.

neuroscience↗

bioRxiv2021

Input to layer 1 of somatosensory cortex: Local input outweighs long-range and can modulate the activity of layer 1 interneurons

Neocortical Layer (L) 1 has been proposed to be at the center for top-down and bottom-up integration. It is a locus for interactions between long-range inputs, L1 interneurons and apical tuft dendrites of pyramidal neurons. While input to L1 has been studied intensively, the level and effect of input to this layer has still not been completely characterized. Here we examined the input to L1 of mouse somatosensory cortex with retrograde tracing and optogenetics. Our assays reveal that local input to L1 is predominantly from L2/3 and L5 pyramidal neurons and interneurons, and that subtypes of local L5 and L6b neurons project to L1 with different probabilities. Long-range input from sensory-motor cortices to L1 of S1 arose predominantly from L2/3 neurons. Our optogenetic experiments showed that intra-telencephalic L5 pyramidal neurons drive L1 interneurons but have no effect locally on L5 apical tuft dendrites. Dual retrograde tracing revealed that a fraction of local and long-range neurons were both presynaptic to L5 neurons and projected to L1. Our work highlights the prominent role of local inputs to L1 and shows the potential for complex interactions between long-range and local inputs which are both in position to modify the output of somatosensory cortex.

neuroscience↗

bioRxiv2021

Post-encoding modulation of spatial memory by a GABAA-agonist

We investigated the role of the post-encoding period for consolidation of self-centered (egocentric) and world-centered (allocentric) spatial memory in neurologically normal human subjects. We used the GABAA-ergic anesthetic propofol to transiently modulate neural activity during the early stage of spatial memory consolidation. A total of 52 patients undergoing minor surgery learned to navigate to a target in a five-armed maze derived from animal experiments immediately prior to injection of propofol (early group) or more than 60 minutes before injection (late group). Two hundred and forty minutes after anesthesia, subjects were tested for memory-guided navigation. Our results show a selective impairment of memory-guided navigation in the early group and near-normal performance in the late group. Both egocentric and allocentric navigation were affected, albeit with distinct error patterns. In the egocentric condition, early group patients navigated significantly more often to a wrong alley of the maze but showed normal navigation times, thus suggesting a deficit mainly for memory of sequences of path segments. By contrast, in the allocentric condition, early group patients mostly navigated to the correct alley of the maze but showed a significant increase in detours and prolonged navigation times, thus suggesting a weakened representation of the relationship between landmarks. We conclude that presumably hippocampus-dependent networks contribute to early consolidation of representations underlying both egocentric and allocentric memory-guided navigation. Distinct aspects of these representations are susceptible to GABAA-ergic modulation within a post-encoding time-window of less than 60 minutes, indicating a redistribution and reconfiguration of spatial memory networks early during consolidation. Significance statementO_LIPropofol modulates consolidation of spatial representations underlying human spatial navigation. C_LIO_LIFollowing administration of propofol, memory-guided navigation using self-centered (egocentric) and world-centered (allocentric) spatial information is impaired. C_LIO_LIError patterns after administration of propofol suggest modulation of a post-encoding integration process relevant for ego- and allocentric memory representations. C_LIO_LIThe transient susceptibility of this process to GABAA-ergic modulation is consistent with rapid reconfiguration of networks for spatial memory shortly after learning. C_LIO_LIPropofol provides a pharmacological tool to investigate spatial memory consolidation in humans. C_LI

neuroscience↗

bioRxiv2021

Double-μPeriscope: a tool for multi-layer optical recordings, optogenetic stimulations or both

Intelligent behavior and cognitive functions in mammals depend on cortical microcircuits made up of a variety of excitatory and inhibitory cells that form a forest-like complex across six layers. Mechanistic understanding of cortical microcircuits requires both manipulation and monitoring of multiple layers and interactions between them. However, existing techniques are limited as to simultaneous monitoring and stimulation at different depths without damaging a large volume of cortical tissue. Here, we present a relatively simple and versatile method for delivering light to any two cortical layers simultaneously. The method uses a tiny optical probe consisting of two micro-prisms mounted on a single shaft. We demonstrate the versatility of the probe in three sets of experiments: first, two distinct cortical layers were optogenetically and independently manipulated; second, one layer was stimulated while the activity of another layer was monitored; third, the activity of thalamic axons distributed in two distinct cortical layers were simultaneously monitored in awake mice. Its simple-design, versatility, small-size and low-cost allow the probe to be applied widely to address important biological questions.

neuroscience↗

bioRxiv2019

Dendritic GABAB receptors control nonlinear information transfer along the dendro-somatic axis in layer 5 pyramidal neurons

Dendritic GABAB receptors (GABABRs) mediate a slow form of interhemispheric inhibition. Surprisingly, this inhibition has no detectable effect on the somatic membrane potential of layer 5 pyramidal neurons, whereas the action potential (AP) output is robustly decreased even when the input is proximal to the cell body. To elucidate the underlying mechanisms, we systematically mapped the AP frequency-current (F-I) relationship during dual patch-clamp recordings from soma and apical dendrite. The AP output function was governed by the synergistic interaction between dendritic and somatic compartments as the local input and transfer resistance from dendrite to soma (Kds) depended on the dendritic membrane potential. Thus, Kds doubled at an estimated rate of once per 28.7 mV depolarization due to HCN channel deactivation. In addition, dendritic L-type Ca2+ channels converted individual APs into dendritic Ca2+ spikes causing high-frequency bursts of APs (HFB) during large dendritic depolarization. Activation of dendritic GABABRs greatly reduced both nonlinear mechanisms. While direct block of L-type Ca2+ channels reduced the number of HFBs, K+ channel activation decreased voltage-dependent input and transfer resistances and decreased the AP rate under all conditions. These results highlight the powerful modulation of the input integration in pyramidal neurons by metabotropic receptor-activated K+ channels.

neuroscience↗

bioRxiv2019

Perirhinal input to neocortical layer 1 controls learning

Signals sent back to the neocortex from the hippocampus control the long-term storage of memories in the neocortex1,2, but the cellular mechanisms underlying this process remain elusive. Here, we show that learning is controlled by specific medial-temporal input to neocortical layer 1. To show this we used direct cortical microstimulation detection task that allowed the precise region of learning to be examined and manipulated. Chemogenetically suppressing the last stage of the medial temporal loop, i.e. perirhinal cortex input to neocortical layer 1, profoundly disrupted early memory formation but had no effect on behavior in trained animals. The learning involved the emergence of a small population of layer 5 pyramidal neurons (~10%) with significantly increased firing involving high-frequency bursts of action potentials that were also blocked by suppression of perirhinal input. Moreover, we found that dendritic excitability was correspondingly enhanced in a similarly-sized population of pyramidal neurons and suppression of dendritic activity via optogenetic activation of dendrite-targeting inhibitory neurons also suppressed learning. Finally, single-cell stimulation of cortical layer 5 pyramidal neurons showed that burst but not regular firing retrieved previously learned behavior. We conclude that the medial temporal input to the neocortex controls learning through a process in L1 that elevates dendritic calcium and promotes burst firing.

neuroscience↗

bioRxiv2019

Whisking signals motor preparation and the behavioral state of mice

A central function of the brain is to plan, predict and imagine the effect of movement in a dynamically changing environment. Here we show that in mice head fixed in a plus-maze, floating on air, and trained to pick lanes based on visual stimuli, the asymmetric movement and position of whiskers on the two sides of the face signals whether the animal is moving, turning, expecting reward or licking. We show that 1) we can decode and predict the behavioral state of the animal based on this asymmetry, 2) that tactile input from whiskers indicates little about the behavioral state, and 3) that movement of the nose correlates with asymmetry, indicating that facial expression of the mouse is itself correlated with behavioral state. Amazingly, the movement of whiskers - a behavior that is not instructed or necessary in the task--informs an observer about what a mouse is doing in the maze. Thus, these mobile tactile sensors reflect a behavioral and movement-preparation state of the mouse.

neuroscience↗