What does the mean mean? A simple test for neuroscience.

Trial-averaged metrics, e.g. tuning curves or population response vectors, are a ubiquitous way of characterizing neuronal activity. But how relevant are such trial-averaged responses to neuronal computation itself? Here we present a simple test to estimate whether average responses reflect aspects of neuronal activity that contribute to neuronal processing. The test probes two assumptions implicitly made whenever average metrics are treated as meaningful representations of neuronal activity: Reliability: Neuronal responses repeat consistently enough across trials that they convey a recognizable reflection of the average response to downstream regions.Behavioural relevance: If a single-trial response is more similar to the average template, it is more likely to evoke correct behavioural responses. We apply this test to two data sets: (1) Two-photon recordings in primary somatosensory cortices (S1 and S2) of mice trained to detect optogenetic stimulation in S1; and (2) Electrophysiological recordings from 71 brain areas in mice performing a contrast discrimination task. Under the highly controlled settings of Data set 1, both assumptions were largely fulfilled. In contrast, the less restrictive paradigm of Data set 2 met neither assumption. Simulations predict that the larger diversity of neuronal response preferences, rather than higher cross-trial reliability, drives the better performance of Data set 1. We conclude that when behaviour is less tightly restricted, average responses do not seem particularly relevant to neuronal computation, potentially because information is encoded more dynamically. Most importantly, we encourage researchers to apply this simple test of computational relevance whenever using trial-averaged neuronal metrics, in order to gauge how representative cross-trial averages are in a given context.

Structural Degradation in Midcingulate Cortex Is Associated with Pathological Aggression in Mice.

Pathological aggression is a debilitating feature of many neuropsychiatric disorders, and cingulate cortex is one of the brain areas centrally implicated in its control. Here we explore the specific role of midcingulate cortex (MCC) in the development of pathological aggression. To this end, we investigated the structural and functional degeneration of MCC in the BALB/cJ strain, a mouse model for pathological aggression. Compared to control animals from the BALB/cByJ strain, BALB/cJ mice expressed consistently heightened levels of aggression, as assessed by the resident-intruder test. At the same time, immunohistochemistry demonstrated stark structural degradation in the MCC of aggressive BALB/cJ mice: Decreased neuron density and widespread neuron death were accompanied by increased microglia and astroglia concentrations and reactive astrogliosis. cFos staining indicated that this degradation had functional consequences: MCC activity did not differ between BALB/cJ and BALB/cByJ mice at baseline, but unlike BALB/cByJ mice, BALB/cJ mice failed to activate MCC during resident-intruder encounters. This suggests that structural and functional impairments of MCC, triggered by neuronal degeneration, may be one of the drivers of pathological aggression in mice, highlighting MCC as a potential key area for pathologies of aggression in humans.

A central role for anterior cingulate cortex in the control of pathological aggression.

Controlling aggression is a crucial skill in social species like rodents and humans and has been associated with anterior cingulate cortex (ACC). Here, we directly link the failed regulation of aggression in BALB/cJ mice to ACC hypofunction. We first show that ACC in BALB/cJ mice is structurally degraded: neuron density is decreased, with pervasive neuron death and reactive astroglia. Gene-set enrichment analysis suggested that this process is driven by neuronal degeneration, which then triggers toxic astrogliosis. cFos expression across ACC indicated functional consequences: during aggressive encounters, ACC was engaged in control mice, but not BALB/cJ mice. Chemogenetically activating ACC during aggressive encounters drastically suppressed pathological aggression but left species-typical aggression intact. The network effects of our chemogenetic perturbation suggest that this behavioral rescue is mediated by suppression of amygdala and hypothalamus and activation of mediodorsal thalamus. Together, these findings highlight the central role of ACC in curbing pathological aggression.

Where is Cingulate Cortex? A Cross-Species View.

To compare findings across species, neuroscience relies on cross-species homologies, particularly in terms of brain areas. For cingulate cortex, a structure implicated in behavioural adaptation and control, a homologous definition across mammals is available - but currently not employed by most rodent researchers. The standard partitioning of rodent cingulate cortex is inconsistent with that in any other model species, including humans. Reviewing the existing literature, we show that the homologous definition better aligns results of rodent studies with those of other species, and reveals a clearer structural and functional organisation within rodent cingulate cortex itself. Based on these insights, we call for widespread adoption of the homologous nomenclature, and reinterpretation of previous studies originally based on the nonhomologous partitioning of rodent cingulate cortex.

Gradient of Parvalbumin- and Somatostatin-Expressing Interneurons Across Cingulate Cortex Is Differentially Linked to Aggression and Sociability in BALB/cJ Mice.

Successfully navigating social interactions requires the precise and balanced integration of social and environmental cues. When such flexible information integration fails, maladaptive behavioral patterns arise, including excessive aggression, empathy deficits, and social withdrawal, as seen in disorders such as conduct disorder and autism spectrum disorder. One of the main hubs for the context-dependent regulation of behavior is cingulate cortex, specifically anterior cingulate cortex (ACC) and midcingulate cortex (MCC). While volumetric abnormalities of ACC and MCC have been demonstrated in patients, little is known about the exact structural changes responsible for the dysregulation of behaviors such as aggression and social withdrawal. Here, we demonstrate that the distribution of parvalbumin (PV) and somatostatin (SOM) interneurons across ACC and MCC differentially predicts aggression and social withdrawal in BALB/cJ mice. BALB/cJ mice were phenotyped for their social behavior (three-chamber task) and aggression (resident-intruder task) compared to control (BALB/cByJ) mice. In line with previous studies, BALB/cJ mice behaved more aggressively than controls. The three-chamber task revealed two sub-groups of highly-sociable versus less-sociable BALB/cJ mice. Highly-sociable BALB/cJ mice were as aggressive as the less-sociable group-in fact, they committed more acts of socially acceptable aggression (threats and harmless bites). PV and SOM immunostaining revealed that a lack of specificity in the distribution of SOM and PV interneurons across cingulate cortex coincided with social withdrawal: both control mice and highly-sociable BALB/cJ mice showed a differential distribution of PV and SOM interneurons across the sub-areas of cingulate cortex, while for less-sociable BALB/cJ mice, the distributions were near-flat. In contrast, both highly-sociable and less-sociable BALB/cJ mice had a decreased concentration of PV interneurons in MCC compared to controls, which was therefore linked to aggressive behavior. Together, these results suggest that the dynamic balance of excitatory and inhibitory activity across ACC and MCC shapes both social and aggressive behavior.

The Virtual-Environment-Foraging Task enables rapid training and single-trial metrics of rule acquisition and reversal in head-fixed mice.

Behavioural flexibility is an essential survival skill, yet our understanding of its neuronal substrates is still limited. While mouse research offers unique tools to dissect the neuronal circuits involved, the measurement of flexible behaviour in mice often suffers from long training times, poor experimental control, and temporally imprecise binary (hit/miss) performance readouts. Here we present a virtual-environment task for mice that tackles these limitations. It offers fast training of vision-based rule reversals (~100 trials per reversal) with full stimulus control and continuous behavioural readouts. By generating multiple non-binary performance metrics per trial, it provides single-trial estimates not only of response accuracy and speed, but also of underlying processes like choice certainty and alertness (discussed in detail in a companion paper). Based on these metrics, we show that mice can predict new task rules long before they are able to execute them, and that this delay varies across animals. We also provide and validate single-trial estimates of whether an error was committed with or without awareness of the task rule. By tracking in unprecedented detail the cognitive dynamics underlying flexible behaviour, this task enables new investigations into the neuronal interactions that shape behavioural flexibility moment by moment.

Relating constructs of attention and working memory to social withdrawal in Alzheimer's disease and schizophrenia: issues regarding paradigm selection.

Central nervous system diseases are not currently diagnosed based on knowledge of biological mechanisms underlying their symptoms. Greater understanding may be offered through an agnostic approach to traditional disease categories, where learning more about shared biological mechanisms across conditions could potentially reclassify sub-groups of patients to allow realisation of more effective treatments. This review represents the output of the collaborative group "PRISM", tasked with considering assay choices for assessment of attention and working memory in a transdiagnostic cohort of Alzheimer's disease and schizophrenia patients exhibiting symptomatic spectra of social withdrawal. A multidimensional analysis of this nature has not been previously attempted. Nominated assays (continuous performance test III, attention network test, digit symbol substitution, N-back, complex span, spatial navigation in a virtual environment) reflected a necessary compromise between the need for broad assessment of the neuropsychological constructs in question with several pragmatic criteria: patient burden, compatibility with neurophysiologic measures and availability of preclinical homologues.

The Virtual-Environment-Foraging Task enables rapid training and single-trial metrics of attention in head-fixed mice.

Attention - the flexible allocation of processing resources based on behavioural demands - is essential to survival. Mouse research offers unique tools to dissect the underlying pathways, but is hampered by the difficulty of accurately measuring attention in mice. Current attention tasks for mice face several limitations: Binary (hit/miss), temporally imprecise metrics, behavioural confounds and overtraining. Thus, despite the increasing scope of neuronal population measurements, insights are limited without equally precise behavioural measures. Here we present a virtual-environment task for head-fixed mice based on 'foraging-like' navigation. The task requires animals to discriminate gratings at orientation differences from 90° to 5°, and can be learned in only 3-5 sessions (<550 trials). It yields single-trial, non-binary metrics of response speed and accuracy, which generate secondary metrics of choice certainty, visual acuity, and most importantly, of sustained and cued attention - two attentional components studied extensively in humans. This allows us to examine single-trial dynamics of attention in mice, independently of confounds like rule learning. With this approach, we show that C57/BL6 mice have better visual acuity than previously measured, that they rhythmically alternate between states of high and low alertness, and that they can be prompted to adopt different performance strategies using minute changes in reward contingencies.

Synchrony makes neurons fire in sequence, and stimulus properties determine who is ahead.

The synchronized activity of cortical neurons often features spike delays of several milliseconds. Usually, these delays are considered too small to play a role in cortical computations. Here, we use simultaneous recordings of spiking activity from up to 12 neurons to show that, in the cat visual cortex, the pairwise delays between neurons form a preferred order of spiking, called firing sequence. This sequence spans up to ∼ 15 ms and is referenced not to external events but to the internal cortical activity (e.g., beta/gamma oscillations). Most importantly, the preferred sequence of firing changed consistently as a function of stimulus properties. During beta/gamma oscillations, the reliability of firing sequences increased and approached that of firing rates. This suggests that, in the visual system, short-lived spatiotemporal patterns of spiking defined by consistent delays in synchronized activity occur with sufficient reliability to complement firing rates as a neuronal code.

Timescales of multineuronal activity patterns reflect temporal structure of visual stimuli.

The investigation of distributed coding across multiple neurons in the cortex remains to this date a challenge. Our current understanding of collective encoding of information and the relevant timescales is still limited. Most results are restricted to disparate timescales, focused on either very fast, e.g., spike-synchrony, or slow timescales, e.g., firing rate. Here, we investigated systematically multineuronal activity patterns evolving on different timescales, spanning the whole range from spike-synchrony to mean firing rate. Using multi-electrode recordings from cat visual cortex, we show that cortical responses can be described as trajectories in a high-dimensional pattern space. Patterns evolve on a continuum of coexisting timescales that strongly relate to the temporal properties of stimuli. Timescales consistent with the time constants of neuronal membranes and fast synaptic transmission (5-20 ms) play a particularly salient role in encoding a large amount of stimulus-related information. Thus, to faithfully encode the properties of visual stimuli the brain engages multiple neurons into activity patterns evolving on multiple timescales.

Neuronal avalanches in spontaneous activity in vivo.

Many complex systems give rise to events that are clustered in space and time, thereby establishing a correlation structure that is governed by power law statistics. In the cortex, such clusters of activity, called "neuronal avalanches," were recently found in local field potentials (LFPs) of spontaneous activity in acute cortex slices, slice cultures, the developing cortex of the anesthetized rat, and premotor and motor cortex of awake monkeys. At present, it is unclear whether neuronal avalanches also exist in the spontaneous LFPs and spike activity in vivo in sensory areas of the mature brain. To address this question, we recorded spontaneous LFPs and extracellular spiking activity with multiple 4 × 4 microelectrode arrays (Michigan Probes) in area 17 of adult cats under anesthesia. A cluster of events was defined as a consecutive sequence of time bins Δt (1-32 ms), each containing at least one LFP event or spike anywhere on the array. LFP cluster sizes consistently distributed according to a power law with a slope largely above -1.5. In two thirds of the corresponding experiments, spike clusters also displayed a power law that displayed a slightly steeper slope of -1.8 and was destroyed by subsampling operations. The power law in spike clusters was accompanied with stronger temporal correlations between spiking activities of neurons that spanned longer time periods compared with spike clusters lacking power law statistics. The results suggest that spontaneous activity of the visual cortex under anesthesia has the properties of neuronal avalanches.

Time delays in the beta/gamma cycle operate on the level of individual neurons.

Recent studies suggested that small time delays among synchronized responses can convey information about visual stimuli. We compared these delays across different types of synchronized signals: single-unit action potentials, multi-unit action potentials, and local field potentials obtained with invasive recordings from cat visual cortex and magnetoencephalographic and electroencephalographic signals recorded from the scalp of human beings. In the signals that reflected more localized sources, time delays were larger and more selective for stimulus properties than in the signals that reflected more large-scale sources. The results suggest that a cortical code for stimulus features that exploits time delays operates predominantly across individual neurons rather than across larger anatomical structures such as brain areas.

Frequencies of gamma/beta oscillations are stably tuned to stimulus properties.

Stimulus-induced changes in oscillation frequencies may affect information flow in the brain. We investigated whether the oscillation frequency of spiking activity in cat area 17 changes as a function of the drifting direction of sinusoidal gratings. Oscillation frequencies were tuned to specific drifting directions, such that some directions induced higher oscillation frequencies than others. When activity from the same neurons was recorded at a later time point, the average oscillation frequency with which the neurons responded had also often changed. However, the direction tuning of the neurons' oscillation frequencies remained constant. Thus, while the overall oscillation frequency, across all drift directions, was state-dependent, the relative change in oscillation frequencies induced by stimulus properties was not, the tuning remaining stable.

Measuring sub-millisecond delays in spiking activity with millisecond time-bins.

Recent evidence indicates that sub-millisecond delays in neuronal spiking activity may be relevant for neural coding. Estimates of these delays are usually made from cross-correlation histograms (CCH) binned to 1ms. We investigated the degree to which it is possible to measure delays with sub-millisecond precision when one computes CCHs with bin sizes > or =1ms. To this end, we introduced sub-millisecond shifts into spike trains recorded from cat visual cortex. The bin sizes of 1/2 to 2ms were the most optimal for measuring the artificial shifts, even when detecting shifts smaller than 0.5ms. The results suggest that preferably, one should use CCHs with approximately 1ms binning even when investigating differences in delays considerably smaller than 1ms.

Spatiotemporal structure in large neuronal networks detected from cross-correlation.

The analysis of neuronal information involves the detection of spatiotemporal relations between neuronal discharges. We propose a method that is based on the positions (phase offsets) of the central peaks obtained from pairwise cross-correlation histograms. Data complexity is reduced to a one-dimensional representation by using redundancies in the measured phase offsets such that each unit is assigned a "preferred firing time" relative to the other units in the group. We propose two procedures to examine the applicability of this method to experimental data sets. In addition, we propose methods that help the investigation of dynamical changes in the preferred firing times of the units. All methods are applied to a sample data set obtained from cat visual cortex.