Projects-CAMP 2015

Michael Hausser (University College, London)

• Reproduce directional selectivity in a passive dendritic cable according to Wilfrid Rall’s original model (Rall W. (1964) Theoretical significance of dendritic trees for neuronal input-output relations. In Neural Theory and Modeling, ed. R.F. Reiss. Stanford Univ. Press). This chapter is reproduced in the book containing the collected papers of Rall, called The Theoretical Foundations of Dendritic Function, MIT Press 1994 (http://bit.ly/a7N2H7).

• Suggested steps:

1. Reproduce the results of the original Rall model (see below for guidelines)

2. Make the conditions more realistic:

• add a soma

• use biexponential synaptic conductances

• shorten the cable: at what length does the effect drop below 5%?

3. Reproduce in a pyramidal cell model with and without NMDA conductances

 Guidelines for reproducing Rall's original simulation of directional selectivity:

1. Cable with 10 segments of 1 μm diameter and 0.2 λ each

2. 8 synapses, one on each compartment, from compartments 2-9

3. Synapses are square conductance pulses of 0.25 τ duration

4. synapses are activated sequentially in pairs at 0.25 τ intervals

5. Maximum synaptic conductance equals 1/Rm

6. Rm = 5000 and Ri = 50

7. Record voltage at compartment 1

Rishikesh Narayanan (Indian Institute of Science, Bangalore)

1. Project 1: Hippocampal pyramidal dendrites express constitutively active G-protein coupled inwardly rectifying potassium (GIRK) channels. Identify the roles of GIRK channels in introducing metaplasticity within a BCM-like plasticity framework.

• Step 1: Ignoring the action of G protein coupling, assess the role of constitutively active GIRK channels (various expression levels) in modulating a BCM-like plasticity profile in an excitatory synapse.(Use: http://senselab.med.yale.edu/ModelDB/ShowModel.asp?model=147538 for a conductance-based BCM-like plasticity rule).

• Step 2: Add an inhibitory synapse containing GABA-B receptors. Couple GABA-B receptors to the GIRK channels using G-protein signaling (http://senselab.med.yale.edu/ModelDB/ShowModel.asp?model=18198). How does GIRK-induced metaplasticity change if you consider feedforward inhibition to control the relative activation of excitatory and inhibitory synapses? Perform sensitivity analyses on each of the associated parameters to assess the robustness of your conclusions.

• Project 2: Activity-dependent modulation of GIRK channels accompanying synaptic plasticity has been experimentally identified. Devise calcium-dependent rules for plasticity in GIRK channels to accompany synaptic plasticity within a BCM-like framework, and assess the physiological role of such concurrent calcium-dependent plasticity in synaptic receptors and GIRK channels.

• References for Projects

1: Chen X, Johnston D. Constitutively active G-protein-gated inwardly rectifying K+ channels in dendrites of hippocampal CA1 pyramidal neurons. J Neurosci. 2005 Apr 13;25(15):3787-92.

2: Chung HJ, Ge WP, Qian X, Wiser O, Jan YN, Jan LY. G protein-activated inwardly rectifying potassium channels mediate depotentiation of long-term potentiation.

Proc Natl Acad Sci U S A. 2009 Jan 13;106(2):635-40.

3: Chung HJ, Qian X, Ehlers M, Jan YN, Jan LY. Neuronal activity regulates phosphorylation-dependent surface delivery of G protein-activated inwardly rectifying potassium channels. Proc Natl Acad Sci U S A. 2009 Jan 13;106(2):629-34.

Raghav Rajan (Indian Institute of Science Education and Research, Pune)

• Project 1: Neural mechanisms for generating syntax-specific neuronal activity

The song of adult birds is a learned sequence of acoustic elements (similar to human speech) that the bird produces in a specific order. Adult birds, sing acoustic elements (syllables) in specific sequences and the probabilities of going from one syllable to another are very stable over time. The neural mechanisms underlying syntax are not well understood.

Recent experimental data has shown that neurons in a premotor nucleus HVC, code for specific aspects of sequence by firing specifically before certain syllable-syllable transitions, firing during the start and termination of repeat sequences, etc. Assuming that such syntax-specific activity is generated by network activity within HVC (and not fed into HVC from its inputs), the aim would be to make realistic models of neurons and networks within HVC to examine the origins of such syntax-specific activity. Specifically, one class of basal ganglia projecting neurons in HVC, (HVCX neurons) show activity that appears to code for specific syllable transitions (look at ref. 1 and 2). HVCX neurons also express some interesting channel combinations – they show slow inhibition mediated by GABAB, mGluR II and GIRK channels (Ref. 7). They also possess Ih currents (Ref. 4).

Goal: To make a realistic model of HVCX neurons and explore ways in which you can get syllable specific activity like the ones seen in ref. 1 and 2. How much does input matter and how much do intrinsic properties of cells matter? In other words, can you give the same input repeatedly to the cell and get different responses because there are slow currents that change the responses to successive pulses? Given that these neurons get gradually hyperpolarised as song begins (see ref. 6), would this account for some of the syllable transition specific activity seen in ref. 2?

• Project 2: How to initiate song?

Song bouts typically start with a variable number of repeats of a short sound called an introductory note (IN). These INs could be some kind of calibration process that happens before song sequences start up (Ref. 2). Song sequence generation is thought to be driven by a synaptic chain like activation of neurons (Ref. 6), something like a syn-fire chain.

Given a model for producing song sequences, how do you start it up each time to mimic different song bouts? In the bird, the main motor neurons driving song are the HVCRA neurons that project to RA from premotor nucleus HVC. These neurons are very hyperpolarised at rest and don't have any spontaneous activity. They are more depolarised during singing and appear to gradually go from hyperpolarised to a depolarised state during INs that precede song motifs (see Ref. 6). What circuit mechanisms can drive this slow depolarisation? And if the membrane potential has to reach a particular threshold for song motifs to be initiated, can this account for the variable number of INs (or variable time from start of the first IN to start of song motifs) seen each time the bird starts a bout of song? Ref. 4 has some models for all HVC neurons and Ref. 5 and 8 are experimental studies of connectivity between HVC neurons.

• References:

1. Neural coding of syntactic structure in learned vocalizations in the songbird

Hisataka Fujimoto, Taku Hasegawa and Dai Watanabe, Journal of Neuroscience, July 6, 2011; 31(27):10023 – 10033.

2. Behavioral and neural signatures of readiness to initiate a learned motor sequence

Raghav Rajan and Allison J Doupe, Current Biology, Jan 7, 2013; 23(1):87-93.

3. Generating variable birdsong syllable sequences with branching chain networks in avian, premotor nucleus HVC

Jin DZ, Phys Rev E Stat Nonlin Soft Matter Phys. 2009 Nov; 80(5 Pt 1):051902. Epub 2009 Nov 5

4. Electrophysiological characterization and computational models of HVC neurons in the zebra finch

Daou A, Ross MT, Johnson F, Hyson RL, Bertram R, J Neurophysiol Sep 2013; 110(5):1227-45

5. Synaptic interactions underlying song-selectivity in the avian nucleus HVC revealed by dual, intracellular recordings

Rosen MJ, Mooney R, J. Neurophysiol. 2006 Feb; 95(2):1158-75

6. Support for a synaptic chain model of neuronal sequence generation

Long MA, Jin DZ, Fee MS, Nature Nov 18 2010; 468(7322):394-9

7. Slow synaptic inhibition mediated by metabotropic glutamate receptor activation of GIRK channels

Dutar et al., J. Neurophys. 2000. 84(5): 2284-2290

8. The HVC microcircuit: the synaptic basis for interactions between song motor and vocal plasticity pathways

Richard Mooney and Jonathan Prather

J. Neurosci. 2005.. 25(8): 1952-1964

Arvind Kumar and Sumantra Chattarji

• Project 1: Fear generalisation in Lateral Amygdala

Recently Ghose and Chattarji (2015) have measured spiking activity in the lateral amygdala changes when animals generalise the fear. Their data shows how the distribution of firing rates changes in the presence of weak and strong aversive stimuli. This change in firing rates and their stimulus specificity underlies changes in the strength of the aversive stimuli and plasticity of the local and input synapses. Moreover, it is also possible that the change in the neuron specificity is related to the baseline firing rate of the neuron —- indeed animal that generalise exhibit high firing rates.

There are two sub-projects that can be initialised during CAMP to better understand how aversive stimulus strength changes the neuronal firing rates

Sub-project 1. Analysis of LA spiking activity in fear generalisation

There is previous evidence from the activity in the central amygdala that fear generalisation involves significant changes in the spontaneous activity of the specific neuronal populations (see Ciocchi et al. 2010)

Study the relationship between the pre-conditioning spontaneous firing rate of the neurons and

(a) change in the neuron’s stimulus specificity

(b) post-conditioning spontaneous activity.

Sub-project 2. Study the effect of the plasticity of recurrent and input synapses in changing the stimulus specificity in a network models of Lateral Amygdala

The goal of the this project is to test the minimal model to explain the change in the stimulus selectivity of Lateral Amygdala neurons. TO this end you will use a recurrent network of excitatory and inhibitory neurons. A fraction of the neurons will be tuned to CS+ and CS- neurons and another partially overlapping population will receive the US input. By adjusting the plasticity of the recurrent and input synapses (both US and CS inputs) we can study how the baseline activity of the neurons and stimulus strength interact to change the stimulus specificity of the LA neurons.

• Suggested reading:

Ghose S, Chattarji S (2015) Neuronal encoding of the switch from specific to generalized fear. Nature Neurosci. 18(1):112-123

Ciocchi S et al. (2010) Encoding of conditioned fear in central amygdala inhibitory circuits. Nature 468:277-282

Arvind Kumar

• Project 1: Herry et al. (2008) described that expression of contextual fear and its subsequent extinction is represented in the basal amygdala by two separate neuron populations. Vlachos et al. (2011)  proposed a model that could integrate the contextual signals with the sensory cue to explain the emergence of the so called `fear' and `extinction' neurons. In that model we proposed that convergence of sensory and contextual cue provide necessary conditions to induce potentiation of the sensory synapses in BA on the neurons which will become '`extinction neurons'. This increases the recurrent inhibition which reduces the activity of `fear' neurons (cf. Vlachos et al. 2011).

Alternatively similar silencing of the `fear' neurons could also be achieved by specifically strengthening of certain specific inhibitory synapses.  Recent data in fact, suggests that there is potentiation of inhibitory synapses in extinction training. (cf. Trouche et al. 2013).

The goal of this project is to reproduce the model proposed in (Vlachos et al. 2011) and study how introduction of inhibitory synaptic plasticity affects the emergence and stability of the fear and extinction neurons.

Suggested reading:

Herry C, Ciocchi S, Senn V, Demmou L, Muller C, et al. (2008) Switching on and off fear by distinct neuronal circuits. Nature 454: 600Ð606.

Vlachos et al.  (2011) Context-dependent encoding of fear and extinction memories in a large-scale network model of the basal amygdala. PLoS Comp. Bio. 7(3): e1001104. doi:10.1371/journal.pcbi.1001104

Trouche et al. (2013) Fear Extinction causes target-specific remodeling of perisomatic inhibitory synapses. Neuron 80 (4), 1054-1065 http://dx.doi.org/10.1016/j.neuron.2013.07.047

Shouval, HZ, Bear, MF, and Cooper, L. N. (2002). A unified model of NMDA receptor-dependent bidirectional synaptic plasticity. Proc. Natl. Acad. Sci. USA 99, 10831-10836.
 

• Project 2: Effect of spike timing and spike plasticity on signal and noise correlations

Correlations in the spiking activity are detrimental to the firing rate based population decoding. Usually from the stimulus decoding perspective two types of correlations are define — signal and noise correlations. Decoding problem becomes simple when +ve signal correlations are accompanied with -ve noise correlations and vice versa (Averbeck et al. 2006). The sign of signal correlations depends on the tuning preferences — neurons with small difference in their preferred orientation have +ve signal correlations. Neurons with similar tuning preferences could be physically clustered (like in cats) or uniformly distributed e.g. in rats. It is poorly understood how does the spatial distribution and recurrent connectivity of the neurons affects the noise correlations in the network.

In project, you can use the standard ring model to study the effect of exc. and inh. STDP (Morrison et al. 2008, Vogels et al. 2011) on the noise correlations in neurons with similar and dissimilar tuning properties (see data in Hansen et al. 2012, Smith and Kohn 2008).

Suggested reading:

Averbeck BB, Latham PE, Pouget A (2006) Neural correlations, population coding and computation. Nature Rev. Neurosci. 7(5) 358-366

Hansen BJ, Chelaru MI, Dragoi V (2012) Correlated variability in laminar cortical circuits. Neuron 76(3): 590-602

Morrison A, Diesmann M, Gerstner W (2008) Phenomenological Models of Synaptic Plasticity based on Spike-Timing. Biological Cybernetics 98(6):459-478. DOI:10.1007/s00422-008-0233-1

Smith M, Kohn A (2008) Spatial and temporal scales of neuronal correlation in primary visual cortex. J. Neurosc. 28(48) 12591-12603

Vogels et al. (2011) Inhibitory plasticity balances excitation and inhibition in sensory pathways and memory networks. Science 334:1569-73. doi: 10.1126/science.1211095

• Project 3: Effect of spike timing and spike rate on the malleability of synapses.

Implement the calcium dependent plasticity rule (or any other rule that you are more familiar with). Pair the activity of the pre- and post-synaptic neurons by externally injecting spikes at different firing rates and latencies and estimate the change in weight after a finite time interval. Compare the effect of using regular and irregular spike trains. Use Gamma process to generate spike trains with different values of regularity (CV). Finally, study the effect of modulation of the firing rate of the Poisson process (i.e. you will generate pre- and post-synaptic spikes using an inhomogeneous Poisson process. Is there any optimal modulation frequency at which the synapses are the most malleable.)

Suggested reading

Kumar A and Mehta MR (2011) Frequency-dependent changes in NMDAR-dependent synaptic plasticity. Front. Comput. Neurosci. 5:38. doi: 10.3389/fncom.2011.00038

Graupner M and Brunel N (2012), A calcium-based plasticity model explains sensitivity of synaptic changes to spike pattern, rate and dendritic location. Proc Natl Acad Sci USA 109 3991-6

Shouval, HZ, Bear, MF, and Cooper, L. N. (2002). A unified model of NMDA receptor-dependent bidirectional synaptic plasticity. Proc. Natl. Acad. Sci. USA 99, 10831-10836.