NeuroWire trainee virtual conference

Uncovering neural circuits: from cellular mechanisms to computational models.

June 2, 2022

NeuroWire Virtual Conference Program

Opening: 10:00

Session 1: 10:10 – 11.00 AM EST

Synaptic plasticity and circuit development
Chair: Eve Honoré, Dr. Jean-Claude Lacaille Lab, Université de Montréal

1.1 Sensory processing dysregulations as reliable translational biomarkers in SYNGAP1 haploinsufficiency.
Maria Isabel Carreno Munoz, Dr. Di Cristo Lab, CHU Sainte Justine/ University of Montreal

1.2 Role of p75 neurotrophin receptor in cortical parvalbumin-positive GABAergic interneurons plasticity and cognitive flexibility.
Pegah Chehrazi, Dr. Di Cristo Lab, CHU Sainte-Justine

1.3 The role of astrocyte glucocorticoid receptors in stress-induced synaptic plasticity deficits.
Ben Rogers, Dr. Ciaran Murphy-Royal, Centre de recherche du Centre Hospitalier de l’Université de Montréal (CRCHUM)

Session 2: 11:00 – 11.45 AM EST

Synaptic integration rules in cortical circuits
Chair: Suzanne van der Veldt, Dr. Bénédicte Amilhon Lab, CHU Sainte-Justine / Université de Montréal

2.1 Lateral entorhinal cortex inputs modulate hippocampal dendritic excitability by recruiting a local disinhibitory microcircuit
Olesia Bilash, Dr. Jayeeta Basu lab, New York University Neuroscience Institute

2.2 Cholinergic Modulation of Dendritic Integration in Primary Visual Cortex of Mice
Mario Galdamez, Dr. Reimer Lab, Neuroscience Graduate Program, Baylor College of Medicine

2.3 Dendritic channelopathies alter the synaptic integration of excitatory inputs in the basal dendrites of layer 5 pyramidal in a mouse model of Fragile X Syndrome. Diana Mitchell, Dr. Roberto Araya’s Lab; CHU Ste-Justine Research Center

Break

Session 3: 12:30 – 13:15 PM EST

Neuronal dynamics underlying memory and behavioral states
Chair: Justine Fortin-Houde, Dr. Bénédicte Amilhon Lab, CHU Sainte-Justine / Université de Montréal

3.1 Utilizing fluorescent neuromodulator reporters to predict cortical acetylcholine and norepinephrine levels from arousal state-related behavioral variables.
Erin Neyhart, Dr. Jacob Reimer Lab, Baylor College of Medicine

3.2 Claustrum lesions lead to changes in behavioural strategy during reversal learning in a spatial memory task.
Vanessa Cattaud, Dr. Jesse Jackson Lab, University of Alberta, Physiology department

3.3 Combination of Dasatinib and Quercetin improves working spatial memory in aged Wistar rats.
Grégory Petrazzo, Laboratory of Cell Biophysics & Laboratory of Molecular bases of Aging at Nencki Institute of Experimental Biology, Polish Academy of Science

Session 4: 13:15 – 14:00 PM EST

Computational circuit modelling in health and disease
Chair: Suhel Tamboli, Dr. Lisa Topolnik Lab, Laval University

4.1 Role of cortical temporal and predictive codes during learning.
Guillaume Etter, Architectures of Biological Learning Lab, Université de Montréal / MILA / McGill

4.2 Age-dependent increased sag current in human pyramidal neurons dampens baseline cortical activity,
Alexandre Guet-McCreight, Dr. Etay Hay Lab – Centre for Addiction and Mental Health, Krembil Centre for Neuroinformatics

4.3 Reduced inhibition in depression impairs stimulus processing in human cortical microcircuits, Heng
Kang Yao, Dr. Etay Hay Lab, Krembil Centre for Neuroinformatics, Centre for Addiction and Mental Health; Department of Physiology, University of Toronto

NeuroWire Virtual Conference Day

Abstracts

Session 1: Synaptic plasticity and circuit development

1.1 Sensory processing dysregulations as reliable translational biomarkers in SYNGAP1 haploinsufficiency
Maria Isabel Carreno Munoz, Dr. Di Cristo Lab. CHU Sainte Justine/ University of Montreal

Amongst the numerous genes associated with intellectual disability, SYNGAP1 stands out for its frequency and penetrance of loss-of-function variants found in patients, as well as the wide range of co-morbid disorders associated with its mutation. Most studies exploring the pathophysiological alterations caused by Syngap1 haploinsufficiency in mouse models have focused on cognitive problems and epilepsy, however whether and to what extent sensory perception and processing are altered by Syngap1 haploinsufficiency is less clear. By performing EEG recordings in awake mice, we identified specific alterations in multiple aspects of auditory and visual processing, including increased baseline gamma oscillation power, increased theta/gamma phase amplitude coupling following stimulus presentation and abnormal neural entrainment in response to different sensory modality-specific frequencies. We also report lack of habituation to repetitive auditory stimuli and abnormal deviant sound detection. Interestingly, we found that most of these alterations are present in human patients as well, thus making them strong candidates as translational biomarkers of sensory-processing alterations associated with SYNGAP1/Syngap1 haploinsufficiency.

1.2 Role of p75 neurotrophin receptor in cortical parvalbumin-positive GABAergic interneurons plasticity and cognitive flexibility
Pegah Chehrazi, Dr. Di Cristo Lab, CHU Sainte-Justine, Université de Montréal

Parvalbumin (PV)-positive GABAergic interneurons constitute the majority of interneurons in the cortex and play a key role in the function and synchronization of cortical networks. Alterations in PV-interneuron connectivity, especially in the medial prefrontal cortex (mPFC), has been found in different psychiatric disorders. We have previously shown that the expression of the p75 neurotrophin receptor (p75NTR) regulates the time course of PV cell synapse maturation in a cell-autonomous fashion. Here, we show that p75NTR removal in postnatal PV cells affects the connectivity and plasticity of PV interneurons and mPFC function in adult Pv-Cre dependent conditional knockout (cKO) mice. We first analyzed the effect of p75NTR removal on the efferent connectivity of PV cells in the perisomatic region and observed that the density of PV+gephyrin+ puncta were significantly increased in mPFC but not in visual cortex (VC) of cKO mice. An important indication of PV cell maturation is the appearance of specialized extracellular matrix structures called perineural nets (PNNs) around the mature cortical PV soma. p75 cKO mouse showed a significant increase in the number of PV cells encircled by PNN, in mPFC but not VC. These mice also showed deficits in attentional set-shifting task, and in extinction of fear memories, both of which rely on mPFC function. We also observed that AAV-mediated reintroduction of P75NTR in PV cells rescues the PV cell efferent connectivity and PNN formation in cKO mice. These findings suggest that P75NTR is a critical regulator of PV cell plasticity, thus modulating cognitive functions, in adult mPFC.

1.3 The role of astrocyte glucocorticoid receptors in stress-induced synaptic plasticity deficits
Ben Rogers, Dr. Ciaran Murphy-Royal Lab, Centre de recherche du Centre Hospitalier de l’Université de Montréal (CRCHUM)

Glucocorticoid receptors (GRs) are key elements in the central response to stress. The predominant view is that stress hormones (i.e. CORT) activate neuronal receptors to elicit synaptic effects. Recent evidence challenges this idea suggesting astrocytes mediate the effects of stress on synapses; however, the precise mechanism by which stress affects astrocytes (i.e. via direct astrocyte GR activation) remains unknown. To investigate this, GRs in astrocytes will be genetically ablated in mice, and subsequently, mice will be exposed to an acute swim-stress paradigm. First, we report that acute swim stress, which we show impairs LTP, correlates with alterations in proteins associated with the regulation of synaptic function. Currently, we are carrying out experiments in GR-flox mice injected with AVV2/5-GfaABC1D-cre into the hippocampus to evaluate whether astrocyte-specific KO of GRs can prevent the impact of stress on these cells and occlude synaptic dysfunction. These data will elaborate on the role of astrocyte GR signalling in stress-induced synaptic dysfunction, further emphasizing the pivotal role of neuronal-glial interactions in mediating the central effects of stress.

Session 2: Synaptic integration rules in cortical circuits

2.1 Lateral entorhinal cortex inputs modulate hippocampal dendritic excitability by recruiting a local disinhibitory microcircuit
Olesia Bilash, Dr. Jayeeta Basu lab, New York University Neuroscience Institute

Episodic memory formation relies on the functional interactions between the entorhinal cortex and hippocampus. Direct projections from lateral entorhinal cortex (LEC) provide the hippocampus with contextual sensory information. Recent studies demonstrated the role of LEC in olfactory, contextual, novelty, and object-related learning, but the circuit interactions between LEC and the hippocampus remain underexplored. We combined functional circuit mapping and computational modeling to examine how long-range glutamatergic LEC inputs modulate the activity of hippocampal area CA1. We demonstrated that these inputs drive excitation and feed-forward inhibition onto CA1 PNs. Using dendritic recordings, we found that photostimulation of LEC axons can drive local dendritic spikes in CA1 PN distal dendrites. To elucidate the circuit mechanisms underlying these dendritic spikes, we probed the GABAergic microcircuit elements recruited by LEC inputs. Using targeted recordings, we found that LEC inputs recruit VIP INs and CCK INs in area CA1. Computational modeling predicted that a VIP IN-mediated disinhibitory microcircuit gates LEC-driven dendritic spikes. Optogenetic silencing of VIP INs confirmed that this is indeed the case. Together, our findings demonstrate that a local disinhibitory microcircuit gates cortically-driven dendritic spikes in area CA1. Our findings suggest that LEC plays an important role in promoting supralinear dendritic computations, likely enabling it to influence plasticity rules and long-term representation in the hippocampus.

2.2 Cholinergic Modulation of Dendritic Integration in Primary Visual Cortex of Mice
Mario Galdamez, Dr. Reimer Lab, Neuroscience Graduate Program, Baylor College of Medicine

Electrophysiological and anatomical studies have shown that feed-forward axon terminals carrying sensory information from the periphery predominantly project to the proximal dendrites of Layer 5 (L5) pyramidal neurons, whereas feedback axon terminals from other cortical regions predominantly project to their apical tufts in Layer 1 (L1). Additionally, L1 contains dense axon terminals from the cholinergic basal forebrain. Previous in vitro studies have shown that acetylcholine (ACh) modulates the excitability of L1 apical dendrites and recent in vivo studies have demonstrated strong correlations between acetylcholine and brain states. However, the relationship between natural fluctuations in acetylcholine levels and dendritic activity in vivo remains unknown. We employed two-photon random access mesoscope (2P-RAM) imaging to simultaneously record acetylcholine fluctuations and neuronal activity of sparsely labeled L5 somata and their apical dendritic arbors in L1, by leveraging the optical fluorescent acetylcholine sensor, AAV-hSyn-GACh3.0, and genetically encoded calcium indicator, AAV-hSyn-FLEX-jGCaMP8s, in primary visual cortex (V1) of awake, behaving mice. Mice were injected in adjacent non-overlapping regions of V1 with the GACh3.0 sensor and a cocktail of diluted AAV-CaMKIIα-Cre (1:20,000) with AAV-FLEX-jGCaMP8s in L5 (500-600 µm depth). With 2P RAM imaging we can manually segment with subcellular resolution functional imaging planes (5-7 planes of 1 px/µm resolution at 5-10 Hz imaging frame rate) traversing the somatodendritic axis of sparsely jGCaMP8s labeled L5 neurons. Our preliminary results from L5 V1 neurons are consistent with recent findings reporting high somatodendritic coupling of L5 pyramidal neurons. By taking advantage of 2P RAM imaging, optical viral constructs, and our computational workflow we have established a pipeline that allows us to characterize how dendritic activity in V1 L5 neurons is modulated by local acetylcholine fluctuations, and generally how dendritic integration might be modulated across different brain states and by different neuromodulators.

2.3 Dendritic channelopathies alter the synaptic integration of excitatory inputs in the basal dendrites of layer 5 pyramidal in a mouse model of Fragile X Syndrome
Diana Mitchell, Dr. Roberto Araya’s Lab, CHU Ste-Justine Research Center, Université de Montréal

Fragile X syndrome (FXS) is the most frequent form of inherited intellectual disability and common known cause of autism spectrum disorders (ASD). Defects in the processing and integration of excitatory inputs in cortical neurons likely contributes to the behavioral phenotype associated with FXS. Layer 5 (L5) pyramidal neurons integrate sensory inputs onto their basal dendrites with information from other cortical areas at their distal dendrites. Here, we aimed to uncover how L5 pyramidal neurons from Fmr1KO mice integrate synaptic inputs at the level of single spines in the basal dendrites. We used two-photon uncaging of caged glutamate to activate nearly simultaneously two clustered spines in L5 pyramidal neurons. While excitatory inputs onto spines integrate linearly before the generation of a dendritic spike in wild-type animals, surprisingly those of Fmr1KO mice summate sublinearly, contradicting what would be expected of a hyperexcitable cortex typically associated with ASD. Since FXS is characterized by several channelopathies in pyramidal cells, we are currently investigating the role of calcium-activated potassium channels in explaining the observed integration defects using genetic manipulations and numerical simulations. Taken together, the results from these experiments will help uncover the role of ion channels in excitatory input integration and identify novel targets for the design of specific drugs to successfully treat FXS. This work was funded by the CIHR, as well as FRQS and QART postdoctoral fellowships to DEM.

Session 3: Neuronal dynamics underlying memory and behavioral states

3.1 Utilizing fluorescent neuromodulator reporters to predict cortical acetylcholine and norepinephrine levels from arousal state-related behavioral variables.
Erin Neyhart, Dr. Jacob Reimer Lab, Baylor College of Medicine

Spontaneous transitions between high and low arousal, indexed by behavioral variables like pupil size and locomotion, have been indirectly linked to cortical levels of acetylcholine (ACh) and norepinephrine (NE). However, we don’t yet know how long ACh & NE are available to act in the cortex and whether this availability is uniform during spontaneous state transitions. The current study used in vivo 2-photon imaging to record ACh and NE activity in various cortical areas of mice using the fluorescent neuromodulator reporters GACh3.0 and GrabNE2h. We first performed controls to validate the sensors in vivo and found that GACh responds faithfully to changes in ACh levels and displays high signal to noise ratio (SNR). We then determined that ACh levels track locomotion and pupil size very closely, and this relationship remains similar across cortical areas. Although GrabNE responds faithfully to NE, its SNR is  lower than that of GACh. We developed a method to reduce noise in the GrabNE signal and determined that cortical NE levels are closely correlated with pupil size, and locomotion-related NE levels remain elevated for longer than ACh levels. Finally, we examined the relationship of cortical ACh and NE to activity of cholinergic basal forebrain (BF) and adrenergic locus coeruleus (LC). Our results indicate that BF and LC both show increased activity directly before locomotion- and dilation-induced peaks in ACh and NE (respectively). From this data we aim to develop a model whereby cortical neuromodulator levels can be predicted from arousal state-related behavior variables.

3.2 Claustrum lesions lead to changes in behavioural strategy during reversal learning in a spatial memory task.
Vanessa Cattaud, University of Alberta, Physiology department

The Claustrum (CLA) is a small subcortical region located between the insula and the putamen. Even though it has been shown that the CLA is directly connected to key structures that regulate memory processing such as prefrontal (PFC), anterior cingulate (ACC), retrosplenial (RSC) and entorhinal (MEC) cortex, its potential role in memory-related mechanisms has yet to be explored. Thus, we aim to investigate whether the CLA is involved in memory processes by focusing on spatial memory. To test this, we specifically lesioned CLA cells projecting to PFC, ACC, RSC and MEC via stereotaxic injections an of apoptosis-inducing virus. Two weeks later, lesioned and control mice were placed on a modified Barnes maze. Our 10-day protocol included 4 days of training, 3 days reversal training and ending with a probe trial 24h following the end of each training sessions. Control and lesioned mice exhibited similar behavior and were capable of spatial learning and spatial memory during the training phase and the probe test, respectively. However, during the reversal training phase, lesioned mice favoured spatial strategy, whereas control mice used preferentially serial strategy. Nevertheless, during the reversal probe trial, there was no significant difference in spatial memory between both groups. To conclude, lesioning the CLA seems to affect behavioural strategy choice while not having a direct impact on spatial memory.

3.3 Combination of Dasatinib and Quercetin improves working spatial memory in aged Wistar rats
Grégory Petrazzo, Laboratory of Cell Biophysics & Laboratory of Molecular bases of Aging at Nencki Institute of Experimental Biology, Polish Academy of Science

Introduction. Cognitive dysfunction negatively impacts the quality-of-life in elders and could sign the onset of dementia. Senescence is a biological process that progressively alter an organ function and it is known to play a central role in mild cognitive impairment. Senolytic drugs have been shown to alleviate symptoms of numerous age-related conditions by reducing the organismal senescent burden. Hypothesis. Combination of Dasatinib and Quercetin (D+Q) senolytics might prevent cognitive decline observed in aged rats.
Objectives. Assess the working spatial memory in aged Wistar rats following D+Q treatment. Determine the long-term efficacy of senolytics use on working spatial memory.
Methods. Young (3-month-old) and naturally aged Wistar male rats (22-month-old) were treated with either D+Q or its vehicle for eight weeks. Before and right after the treatment period, the animals were tested in the active allothetic place avoidance task. Another cohort of aged Wistar male rats (18-month-old) were treated similarly and tested before, right after and six-weeks after the treatment discontinuation.
Results. We confirmed the cognitive decline of aged rats compare to their younger counterpart. We observed in aged but not in young rats treated with D+Q a reduction in memory impairments. Furthermore, D+Q treatment retains long-lasting effects up to six weeks after treatment discontinuation. Conclusion. Our study brings new insights on the effects of D+Q senolytics in alleviating age-associated cognitive dysfunctions.

Session 4: Computational circuit modelling in health and disease

4.1 Role of cortical temporal and predictive codes during learning
Guillaume Etter. Architectures of Biological Learning Lab, Université de Montréal / MILA / McGill

Neurons are the building blocks of computations in the brain. Synapses, the points of contact between neurons, can be enhanced or weakened to fine tune these computations through experience with the ultimate objective to improve behavior. While synaptic plasticity has been well studied in vitro, the large-scale rules governing potentiation in vivo are unknown. Using deep learning models of the neocortex, as well as in vivo calcium imaging and optogenetics in the hippocampus, we explore the neural activity dynamics that underpin learning and memory. We developed a deep learning model of cortical layer 2/3 pyramidal neurons that resolves ambiguous sensory signals arriving at basal dendrites by learning a feedback mapping of internal predictive signals onto apical dendrites. Our model learns to use predictions to resolve ambiguous signals and sheds light on candidate local learning rules in the neocortex. Next, we propose that the hippocampus is instrumental in propagating predictive signals to the neocortex. Using in vivo calcium imaging, we record large neuronal assemblies and disentangle spatiotemporal codes. By optogenetically scrambling temporal input signals to the hippocampus, we show that memory performance, but not spatiotemporal codes are affected. In addition to proposing channels through which neuroscience and artificial intelligence can interact, we discuss candidate mechanisms for learning in the neocortex that depend on top-down, hippocampal predictive signals and could be key in understanding how learning takes place in vivo.

4.2 Age-dependent increased sag current in human pyramidal neurons dampens baseline cortical activity
Alexandre Guet-McCreight, Dr. Etay Hay Lab – Centre for Addiction and Mental Health, Krembil Centre for Neuroinformatics

Aging involves various neurobiological changes, although their effect on brain function in humans remains poorly understood. The growing availability of human neuronal and circuit data provides opportunities for uncovering age-dependent changes of brain networks and for constraining models to predict consequences on brain activity. Here we found increased sag current in human layer 5 pyramidal neurons from older subjects, and captured this effect in biophysical models of younger and older pyramidal neurons. We used these models to simulate detailed layer 5 microcircuits and found lower baseline firing in older pyramidal neuron microcircuits, with minimal effect on response. We then validated the predicted reduced baseline firing using extracellular multi-electrode recordings from human brain slices of different ages. Our results thus report changes in human pyramidal neuron input integration properties and provide fundamental insights on the neuronal mechanisms of altered cortical excitability and resting state activity in human aging.

4.3 Reduced inhibition in depression impairs stimulus processing in human cortical microcircuits
Heng Kang Yao, Etay Hay Lab, Krembil Centre for Neuroinformatics, Centre for Addiction and Mental Health; Department of Physiology, University of Toronto

Reduced inhibition from somatostatin (SST) interneurons is implicated in underlying cognitive deficits in treatment-resistant major depression disorder (depression), but  he link remains to be better established in humans. Here, we tested the effect of reduced somatostatin interneuron-mediated inhibition on cortical processing in human neuronal microcircuits using a data-driven computational approach. We integrated human cellular, circuit, and gene expression data to generate detailed models of human cortical microcircuits in health and depression. We simulated microcircuit baseline and response activity and found a reduced signal-to-noise ratio and increased false/failed detection of stimuli due to a higher baseline activity in depression. We thus applied models of human cortical microcircuits to demonstrate mechanistically how reduced inhibition impairs cortical processing in depression, providing quantitative links between altered inhibition and cognitive deficits.