Biology subjects

Killcross, S.

Publications and source records attributed to Killcross, S..

13 recordsLinked to original sources

bioRxiv2026

Representational geometry of punishment learning across basolateral amygdala neurons

Punishment learning allows animals to suppress actions that cause harm while maintaining actions that earn reward. Although the basolateral amygdala is necessary for punishment learning, how its population activity changes across learning, and how this activity reflects an action's consequences, remain unknown. We used longitudinal calcium imaging to track 1,147 basolateral amygdala neurons in male mice across instrumental training, punishment, choice test, and punishment extinction. Punishment learning transformed basolateral amygdala neuron activity associated with the punished action. This activity diverged from that associated with the unpunished action and became aligned with activity evoked by the punisher itself. This alignment was action and contingency specific. It preserved the direction of individual neurons' responses to the punisher: neurons activated by the punisher become more activated by the punished action alone, whereas neurons inhibited by the punisher become more inhibited by the punished action alone. We then asked whether reward and punishment were encoded separately or on a common value dimension across basolateral amygdala neurons. Punishment added a strong shock component to action-related activity without reducing its reward component, while reward and shock themselves occupied largely orthogonal population directions. Thus, during punishment learning, basolateral amygdala population activity can encode opposing consequences of the same action instead of simply changing what that action is worth.

neuroscience↗

bioRxiv2026

Number, order and time influence how the brain integrates distinct experiences

How does the brain integrate experiences that are separated in time? The present study addressed this question using sensory preconditioning protocols in rats. In these protocols, rats integrate an A-B association (e.g., tone-light) formed in stage 1 with a B-shock association (e.g., light-shock) formed in stage 2 to generate fear responses (freezing) when tested with A alone in stage 3. Here we show that the mechanism of integration depends on the number, order and timing of events across the two stages of training. When the events are novel (low number of exposures) and the interval between stages 1 and 2 is short (24 hours), the A-B and B-shock associations are integrated through formation of a mediated A-shock association during stage 2. By contrast, when the events are more familiar (greater number of exposures), ordered in a particular way, and the interval between stages 1 and 2 is long (14-days), the A-B and B-shock associations are integrated through their chaining at the time of testing with A alone. Thus, number, order and time determine how distinct experiences are integrated in the brain. These findings are discussed with respect to theories of integration and information processing in the medial temporal lobe.

neuroscience↗

bioRxiv2025

The brain architecture of punishment learning

Learning which actions cause harm is essential for survival, yet how the brain transforms this knowledge into adaptive control of behavior remains unclear. We combined whole-brain analysis, multisite chemogenetic silencing, spatial transcriptomics, and longitudinal calcium imaging to map how punishment learning reorganizes brain networks across scales. Learning reshaped mesoscale community organization into a network anchored by the amygdala, subthalamic-hypothalamic zone, and ventral midbrain tegmentum. These regions contributed distinct components of adaptive avoidance and engaged diverse transcriptional programs across multiple neuronal subclasses. Within this network, the amygdala acted as a hub, partitioning action and outcome information and remapping neural geometry to segregate punished from safe actions. Together, these results identify a multiscale neural architecture through which aversive experience can be transformed into flexible behavioral control.

neuroscience↗

bioRxiv2025

The basolateral amygdala complex and perirhinal cortex represent focal and peripheral states of information processing in rats

This study used variations of a sensory preconditioning protocol in male and female rats to test a theory that the basolateral amygdala complex (BLA) and perirhinal cortex (PRh) represent focal and peripheral states of attention, respectively. It specifically tested predictions derived from the theory regarding when learning about a stimulus that signals danger will be disrupted by BLA or PRh infusions of the N-methyl-D-aspartate receptor (NMDAR) antagonist, DAP5. Consistent with the theory, the effects of these infusions depended on the novelty/familiarity of the conditioned stimulus as well as the manner in which it was paired with foot shock. When a stimulus was novel, its conditioning required activation of NMDAR in the BLA and not the PRh (Experiments 2A and 2B) regardless of whether the stimulus-shock pairings were contiguous or separated in time. When a pre-exposed and, thereby, familiar stimulus was presented contiguously with shock, its conditioning again required activation of NMDAR in the BLA and not the PRh (Experiments 1A, 1B, 3A and 3B). However, when a pre-exposed stimulus was indirectly paired with shock - because it was associatively activated at the time of shock or separated from the shock by another stimulus - its conditioning required activation of NMDAR in the PRh and not the BLA (Experiments 1A, 1B, 3A and 3B). These findings are discussed in relation to theories of information processing that distinguish between focal and peripheral states of attention/memory, and past studies that have examined the substrates of learning and memory in the PRh and BLA.

neuroscience↗

bioRxiv2024

Integration of sensory and fear memories in the rat medial temporal lobe

Wong et al. (2019) used a sensory preconditioning protocol to examine how sensory and fear memories are integrated in the rat medial temporal lobe. In this protocol, rats integrate a sound-light (sensory) memory that forms in stage 1 with a light-shock (fear) memory that forms in stage 2 to generate fear responses (freezing) across test presentations of the sound in stage 3. Here, we advance this research by showing that: 1) how/when rats integrate the sound-light and light-shock memories (online in stage 2 or at test in stage 3) changes with the number of sound-light pairings in stage 1; and 2) regardless of how/when it occurs, the integration requires communication between two regions of the medial temporal lobe: the perirhinal cortex and basolateral amygdala complex. Thus, "event familiarity" determines how/when sensory and fear memories are integrated but not the circuitry by which the integration occurs: this remains the same.

neuroscience↗

bioRxiv2024

Prediction error determines how memories are organized in the brain: a study of Pavlovian fear extinction in rats

How is new information organized in memory? According to latent state theories, this is determined by the level of surprise, or prediction error, generated by the new information: small prediction error leads to updating of existing memory, large prediction error leads to encoding of a new memory. We tested this idea using a protocol in which rats were first conditioned to fear a stimulus paired with shock. The stimulus was then gradually extinguished by progressively reducing the shock intensity until the stimulus was presented alone. Consistent with latent state theories, this gradual extinction protocol (small prediction errors) was better than standard extinction (large prediction errors) in producing long-term suppression of fear responses; and the benefit of gradual extinction was due to updating of the conditioning memory with information about extinction. Thus, prediction error determines how new information is organized in memory, and latent state theories adequately describe the ways in which this occurs.

animal behavior and cognition↗

bioRxiv2022

Danger changes the way the brain consolidates neutral information; and does so by interacting with processes involved in the encoding of that information.

This study examined the effect of danger on consolidation of neutral information in two regions of the rat (male and female) medial temporal lobe: the perirhinal cortex (PRh) and basolateral amygdala complex (BLA). The neutral information was the association that forms between an auditory stimulus and a visual stimulus (labelled S2 and S1) across their pairings in sensory preconditioning. We show that, when the sensory preconditioning session is followed by a shocked context exposure, the danger shifts consolidation of the S2-S1 association from the PRh to the BLA; and does so by interacting with processes involved in encoding of the S2-S1 pairings. Specifically, we show that the initial S2-S1 pairing in sensory preconditioning is encoded in the BLA and not the PRh; whereas the later S2-S1 pairings are encoded in the PRh and not the BLA. When the sensory preconditioning session is followed by a context alone exposure, the BLA-dependent trace of the early S2-S1 pairings decays and the PRh-dependent trace of the later S2-S1 pairings is consolidated in memory. However, when the sensory preconditioning session is followed by a shocked context exposure, the PRh-dependent trace of the later S2-S1 pairings is suppressed and the BLA-dependent trace of the initial S2-S1 pairing is consolidated in memory. These findings are discussed with respect to mutually inhibitory interactions between the PRh and BLA, and the way that these regions support memory in other protocols, including recognition memory in people. Significance StatementThe perirhinal cortex (PRh) and basolateral amygdala complex (BLA) process the pairings of neutral auditory and visual stimuli in sensory preconditioning. The involvement of each region in this processing is determined by the novelty/familiarity of the stimuli as well as events that occur immediately after the preconditioning session. Novel stimuli are represented in the BLA; however, as these stimuli are repeatedly presented without consequence, they come to be represented in the PRh. Whether the BLA- or PRh-dependent representation is consolidated in memory depends on what happens next. When nothing of significance occurs, the PRh-dependent representation is consolidated and the BLA-dependent representation decays; but when danger is encountered, the PRh-dependent representation is inhibited and the BLA-dependent representation is selected for consolidation.

animal behavior and cognition↗

bioRxiv2020

The role of the rodent lateral orbitofrontal cortex in simple Pavlovian cue-outcome learning depends on training experience

The orbitofrontal cortex (OFC) is a critical structure in the flexible control of value-based behaviours. OFC dysfunction is typically only detected when task or environmental contingencies change, against a backdrop of apparently intact initial acquisition and behaviour. While intact acquisition following OFC lesions in simple Pavlovian cue-outcome conditioning is often predicted by models of OFC function, this predicted null effect has not been thoroughly investigated. Here we test the effects of lesions and temporary muscimol inactivation of the rodent lateral OFC on the acquisition of a simple single cue-outcome relationship. Surprisingly, pre-training lesions significantly enhanced acquisition after over-training whereas post-training lesions and inactivation significantly impaired acquisition. This impaired acquisition to the cue reflects a disruption of behavioural control and not learning since the cue could also act as an effective blocking stimulus in an associative blocking procedure. These findings suggest that even simple cue-outcome representations acquired in the absence of OFC function are impoverished. Therefore, while OFC function is often associated with flexible behavioural control in complex environments, it is also involved in very simple Pavlovian acquisition where complex cue-outcome relationships are irrelevant to task performance.

neuroscience↗

bioRxiv2020

Medial Orbitofrontal Cortex Regulates Instrumental Conditioned Punishment, but not Pavlovian Conditioned Fear

Bidirectionally aberrant medial orbitofrontal cortical (mOFC) activity has been consistently linked with compulsion and compulsive disorders. Although rodent studies have established a causal link between mOFC excitation and compulsive-like actions, no such link has been made with mOFC inhibition. Here we use excitotoxic lesions of mOFC to investigate its role in sensitivity to punishment; a core characteristic of many compulsive disorders. In our first experiment, we demonstrated that mOFC lesions prevented instrumental conditioned punishment learning in a manner that could not be attributed to differences in Pavlovian conditioned fear. We then showed that increasing the frequency of punishing outcomes allowed mOFC-lesioned animals to overcome their initial deficit. Our second experiment demonstrated that the retrieval of instrumental punishment is also mOFC-dependent, as mOFC lesions prevented the extended retrieval of punishment contingencies relative to shams. In contrast, mOFC lesions did not prevent the re-acquisition of conditioned punishment that was learned prior to lesions being administered. Together, these results reveal that the mOFC does indeed regulate punishment learning and retrieval in a manner that is disassociated from any role in Pavlovian fear learning. These results imply that aberrant mOFC activity may contribute to the punishment insensitivity that is observed across multiple compulsive disorders.

neuroscience↗

bioRxiv2020

The mesolimbic dopamine signatures of relapse to alcohol-seeking

The mesolimbic dopamine system comprises distinct compartments supporting different functions in learning and motivation. Less well understood is how complex addiction-related behaviors emerge from activity patterns across these compartments. Here we show how different forms of relapse to alcohol-seeking are assembled from activity across the ventral tegmental area and the nucleus accumbens. Using gCaMP and dLight fibre photometry, we show that self-administration and two forms of relapse (renewal/context-induced reinstatement and reacquisition) are associated with recruitment across the mesolimbic dopamine system. Using a variety of interventions, we show that this activity is causal to both forms of relapse. Finally, we use dissimilarity matrices to identify mesolimbic dopamine signatures of self-administration, extinction, and relapse. We show that signatures of relapse can be identified from heterogeneous activity profiles across the mesolimbic dopamine system and that these signatures differ for different forms of relapse.

neuroscience↗

bioRxiv2019

Punishment insensitivity emerges from impaired contingency detection, not aversion insensitivity or reward dominance

Our behaviour is shaped by its consequences - we seek rewards and avoid harm. It has been reported that individuals vary markedly in their avoidance of detrimental consequences, i.e. in their sensitivity to punishment. The underpinnings of this variability are poorly understood; they may be driven by differences in aversion sensitivity, motivation for reward, and/or instrumental control. We examined these hypotheses by applying several analysis strategies to the behaviour of rats (n = 48; 18 female) trained in a task permitting concurrent assessment of punishment, reward-seeking, and Pavlovian fear. We show that punishment insensitivity is a unique phenotype, unrelated to differences in reward-seeking and Pavlovian fear, and due to a failure of instrumental control. Subjects insensitive to punishment are afraid of aversive events, they are simply unable to change their behaviour to avoid them.

neuroscience↗

bioRxiv2019

Complementary roles for ventral pallidum cell types and their projections in relapse

Ventral pallidum (VP) is a key node in the neural circuits controlling relapse to drug seeking but how this role relates to different VP cell types and their projections is poorly understood. Using male rats, we show how different forms of relapse to alcohol-seeking are assembled from VP cell types and their projections to lateral hypothalamus (LH) and ventral tegmental area (VTA). First, we used RNAScope in situ hybridization to characterize activity of different VP cell types during relapse to alcohol-seeking provoked by renewal (context-induced reinstatement). We found that VP Gad1 and parvalbumin (PV), but not vGlut2, neurons show relapse-associated changes in c-Fos expression. Next, we used retrograde tracing, chemogenetic, optogenetic, and electrophysiological approaches to study the roles of VPGad1 and VPPV neurons in relapse. We show that VPGad1 neurons contribute to contextual control over relapse (renewal), but not to relapse during reacquisition, via projections to LH where they converge with ventral striatal inputs onto LHGad1 neurons. This convergence of striatopallidal inputs at the level of individual LHGad1 neurons may be critical to balancing propensity for relapse versus abstinence. In contrast, VPPV neurons contribute to relapse during both renewal and reacquisition via projections to VTA but not LH. These findings show complementary roles for different VP cell types and their projections in relapse. VPGad1 neurons control relapse during renewal via projections to LH whereas VPPV neurons control relapse during both renewal and reacquisition via projections to VTA. Targeting these different pathways may provide tailored interventions for different forms of relapse.

neuroscience↗

bioRxiv2018

The rodent lateral orbitofrontal cortex represents expected Pavlovian outcome value but not identity

The orbitofrontal cortex (OFC) is critical for updating reward-directed behaviours flexibly when task contingencies are reversed, or when outcomes are devalued. We systematically examined the generality of these findings using lesions of the rodent lateral OFC (LO) in instrumental action-outcome, and Pavlovian cue-outcome, learning using specific satiety and taste aversion methods of outcome devaluation. LO lesions disrupted outcome devaluation in Pavlovian but not instrumental procedures. Furthermore, this effect was only observed when using taste-aversion devaluation. Using a specific Pavlovian-to-Instrumental transfer procedure, we established that LO is not necessary for the representation of specific outcome properties, but rather in using these properties to access the current motivational value of outcomes. The role of LO in outcome devaluation and reversal learning was also dissociable between anterior and posterior subregions. These novel dissociable task- and subregion-specific effects suggest a way to reconcile contradictory findings between rodent and non-human primate OFC research.

neuroscience↗