* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Short title: Thalamocortical computations during tactile sensation
Neuroethology wikipedia , lookup
Environmental enrichment wikipedia , lookup
Theta model wikipedia , lookup
Neuroplasticity wikipedia , lookup
Recurrent neural network wikipedia , lookup
Electrophysiology wikipedia , lookup
Endocannabinoid system wikipedia , lookup
Artificial general intelligence wikipedia , lookup
Neuroeconomics wikipedia , lookup
Apical dendrite wikipedia , lookup
Axon guidance wikipedia , lookup
Neural modeling fields wikipedia , lookup
Convolutional neural network wikipedia , lookup
Synaptogenesis wikipedia , lookup
Activity-dependent plasticity wikipedia , lookup
Multielectrode array wikipedia , lookup
Single-unit recording wikipedia , lookup
Molecular neuroscience wikipedia , lookup
Neurotransmitter wikipedia , lookup
Types of artificial neural networks wikipedia , lookup
Neural correlates of consciousness wikipedia , lookup
Nonsynaptic plasticity wikipedia , lookup
Clinical neurochemistry wikipedia , lookup
Mirror neuron wikipedia , lookup
Caridoid escape reaction wikipedia , lookup
Development of the nervous system wikipedia , lookup
Neural oscillation wikipedia , lookup
Stimulus (physiology) wikipedia , lookup
Spike-and-wave wikipedia , lookup
Metastability in the brain wikipedia , lookup
Biological neuron model wikipedia , lookup
Neuroanatomy wikipedia , lookup
Chemical synapse wikipedia , lookup
Circumventricular organs wikipedia , lookup
Neuropsychopharmacology wikipedia , lookup
Central pattern generator wikipedia , lookup
Neural coding wikipedia , lookup
Efficient coding hypothesis wikipedia , lookup
Optogenetics wikipedia , lookup
Premovement neuronal activity wikipedia , lookup
Feature detection (nervous system) wikipedia , lookup
Pre-Bötzinger complex wikipedia , lookup
Nervous system network models wikipedia , lookup
Gutnisky et al 11/28/2016 1 1 2 Mechanisms underlying a thalamocortical transformation during active 3 tactile sensation 4 Diego A. Gutnisky1, Jianing Yu1, S. Andrew Hires1,2, Minh-Son To3, Michael Bale4,5, Karel 5 Svoboda1, David Golomb1,6 6 7 Short title: Thalamocortical computations during tactile sensation 8 9 1 10 11 2 12 13 3 14 15 4 16 17 18 6 Janelia Research Campus, HHMI, Ashburn, VA 20147, U.S.A. Dept. of Biological Sciences, Neurobiology Section, University of Southern California, Los Angeles, CA 90089, U.S.A. Dept. of Human Physiology and Centre for Neuroscience, Flinders University,South Australia, Australia. Faculty of Life Sciences, University of Manchester, UK. 5School of Life Sciences, University of Sussex, UK. Depts. of Physiology and Cell Biology and Physics and Zlotowski Center for Neuroscience, Faculty of Health Sciences, Ben Gurion University, Be’er-Sheva 8410501, Israel. ORCID ID: 0000-0001-65753685 19 20 Correspondence should be addressed to [email protected] (K.S.); [email protected] 21 (D.G.) 22 23 24 Author contributions: D.A.G., K.S. and D.G. designed research; D.A.G., J.Y., and S.A.H. 25 performed experiments; D. G., D.A.G. and K.S. contributed to circuit models; D.G. 26 programmed the simulations; D.A.G., D.G. and K.S. analyzed data; D.A.G., K.S. and D.G. 27 wrote the paper. J.Y, S.A.H., M.S.T and M.B. helped with data collection. 28 Gutnisky et al 11/28/2016 2 29 The research was supported by a grant from the United States-Israel Binational Science 30 Foundation (BSF), Jerusalem, Israel No. 2013033 (S.A.H, K.S. and D.G)., the Janelia Research 31 Campus of the Howard Hughes Medical Institute (D.A.G., S.A.H., K.S. and D.G.), the Israel 32 Science Foundation grant No. 88/13 (D.G.), and the Helmsley Charitable Trust through the 33 Agricultural, Biological and Cognitive Robotics Center of Ben-Gurion University of the Negev 34 (D.G.). 35 36 37 Gutnisky et al 11/28/2016 3 38 Abstract 39 During active somatosensation, neural signals expected from movement of the sensors are 40 suppressed in the cortex, whereas information related to touch is enhanced. This tactile 41 suppression underlies the brain’s ability to make fine tactile discriminations. Layer (L) 4 42 excitatory neurons in the barrel cortex, the major target of the somatosensory thalamus (VPM), 43 respond to touch, but have low spike rates and low sensitivity to the movement of whiskers. 44 Most neurons in VPM respond to touch and also show an increase in spike rate with whisker 45 movement. These data imply that signals related to self-movement are suppressed in L4. Fast- 46 spiking (FS) interneurons in L4 respond similar to VPM neurons. Stimulation of halorhodopsin 47 in FS interneurons causes a reduction in FS neuron activity and an increase in L4 excitatory 48 neuron activity. To explain these experimental observations, we constructed a model of the L4 49 circuit, with connectivity constrained by in vitro measurements. The model explores the 50 various synaptic conductance strengths for which L4 FS neurons actively suppress baseline and 51 movement-related activity in layer 4 excitatory neurons. Feedforward inhibition, in concert 52 with recurrent intracortical circuitry, produces tactile suppression. Synaptic delays in 53 feedforward inhibition allow transmission of temporally brief volleys of activity associated 54 with touch. Our model provides a mechanistic explanation of a behavior-related computation 55 implemented by the thalamocortical circuit. 56 Gutnisky et al 11/28/2016 4 57 Author Summary 58 We study how information is transformed between connected brain areas: the thalamus, the 59 gateway to the cortex, and layer 4 (L4) in cortex, which is the first station to process sensory 60 input from the thalamus. When mice perform an active object localization task with their 61 whiskers, thalamic neurons and inhibitory fast-spiking (FS) interneurons in L4 encode whisker 62 movement and touch, whereas L4 excitatory neurons respond almost exclusively to touch. To 63 explain these observations, we constructed a computational model based on measured circuit 64 parameters. The model reveals that without touch, when thalamic activity varies slowly, strong 65 inhibition from FS neurons prevents activity in L4 excitatory neurons. Brief and strong touch- 66 induced thalamic activity excites both excitatory and FS neurons in L4. FS neurons inhibit 67 excitatory neurons with a delay of approximately 1 ms relative to ascending excitation, 68 allowing L4 excitatory neurons to spike. Our results demonstrate that cortical circuits exploit 69 synaptic delays for fast computations. Similar mechanisms likely also operate for rapid stimuli 70 in the visual and auditory systems. Gutnisky et al 11/28/2016 5 71 Introduction 72 Thalamocortical circuits represent model systems for multi-area computations [1]. Sensory 73 information enters the cortex through the thalamus. Transformations in thalamocortical circuits 74 have mostly been studied in anesthetized animals with passive sensory stimuli [2-8] or with 75 artificial whisking [9]. In the somatosensory system these studies have revealed subtle 76 differences in receptive field structure across neurons in the thalamocortical circuit [2-4]. 77 However, active sensation in awake animals involves dynamic interactions with the world, 78 such as saccades [10], palpation with the digits of the hand [11], or movements of the whiskers 79 on the face of rodents [12-14]. During haptic exploration, touch generates ‘exafferent’ signals, 80 while movement itself activates peripheral sensors to produce ‘reafferent’ signals [15-21]. The 81 brain needs to parse these different signals for perception [13]. During active sensation, 82 movement attenuates the transmission of certain sensory signals to the cortex [22-24]. Tactile 83 suppression is thought to enhance perception of salient events that cannot be predicted based 84 on movement. 85 The lemniscal pathway starts with mechanosensory afferents in the whisker follicle. 86 Information then flows through the trigeminal ganglion, to the brainstem, thalamus (barreloids 87 in the ventral posterior medial thalamic nucleus, VPM) and terminates in the primary 88 somatosensory cortex (vS1). The main target of VPM axons is the Layer 4 (L4) barrels in vS1. 89 The microcircuit of each L4 barrel is mostly contained within the barrel, and the connections 90 between specific cell types within the barrel have been mapped: L4 excitatory neurons and L4 91 fast-spiking, parvalbumin (PV)-expressing GABAergic interneurons (FS) are connected within 92 type and across types [25, 26]. Apart from neuromodulation, the only known long-range input Gutnisky et al 11/28/2016 6 93 to L4 originates in VPM [27-29]. VPM excites all L4 neuron types, and L4 FS neurons inhibit 94 the excitatory neurons to implement feedforward inhibition [30-33]. 95 Cell type-specific recordings from VPM, L4 excitatory [34] and L4 FS neurons [35] uncovered 96 a fundamental computation performed by L4 barrels. Neurons in VPM respond to touch, but 97 they also increase their activity during whisker movement (‘whisking’) [17-21]. L4 FS neurons 98 have nearly identical dynamics as VPM neurons. In contrast, L4 excitatory neuron spikes are 99 strongly coupled to touch, but respond only weakly to whisking [34]. L4 microcircuits 100 therefore transmit touch signals and suppress reafferent signals generated by whisking. These 101 observations indicate that the thalamocortical circuit accentuates salient tactile information by 102 suppressing signals related to self-movement. 103 Multiple observations regarding the L4 circuit remain to be explained [35]. First, the circuit 104 suppresses self-movement signals but transmit touch signals. From a theoretical perspective, 105 this selective filtering violates the linear response expected from theories of strongly coupled 106 cortical networks [36, 37]. Second, whereas the baseline spike rates of VPM and L4 FS 107 neurons are substantial during baseline conditions and more than double during whisking, the 108 spike rate of L4 excitatory neurons is very low at baseline and does not increase during 109 whisking. Third, when an inhibitory opsin (halorhodopsin) is excited in L4 FS neurons, the 110 spike rates of FS neurons decrease and the spike rates of L4 excitatory neurons increase [35]. 111 Although this result is naively expected, it contradicts the "paradoxical effect" predicted from 112 circuit models with strong synaptic coupling [38-40]. According to these models, both 113 inhibitory and excitatory neurons should respond with increased spike rates at modest levels of 114 inactivation of inhibitory neurons. Gutnisky et al 11/28/2016 7 115 116 To understand the mechanisms underlying these experimental observations we constructed and 117 analyzed a conductance-based model of the L4 circuit. The synaptic circuits of L4 of the barrel 118 cortex have been studied in great detail in brain slices. These measurements allowed us to 119 constrain the numbers of neurons, neuronal properties, the patterns of connectivity, and the 120 average synaptic strengths in the model. We studied how the model networks responds to 121 thalamic input at baseline, during whisker movement and during touch. We used the model to 122 disentangle the roles of feedforward synaptic connections from the thalamus and recurrent 123 intracortical connections in shaping L4 dynamics. The model revealed an important role for 124 synaptic delays and fast synaptic kinetics in shaping the L4 responses during behavior. 125 126 Results 127 Thalamocortical transformation of tactile information 128 We first summarize recordings made in VPM and in L4 from excitatory and FS GABAergic 129 interneurons in mice performing an object location discrimination behavior (Fig. 1a,b). Head- 130 restrained mice had to localize a vertical pole with their whiskers for a water reward [41, 42]. 131 Whisker movement and touch were tracked on millisecond time scales with high-speed 132 videography [43, 44] (Fig. 1a). Recordings were made with extracellular silicon probe 133 recordings in VPM [35], and loose-seal cell-attached and whole cell recordings in L4 [34, 35, 134 45]. Gutnisky et al 11/28/2016 8 135 Fig. 1. Summary of experimental results in the object-localization task [35]. (a) Top, 136 schematic illustrates measurement of whisker position (azimuthal angle θ), instances of touch 137 and an example trace of whisker position. (b) Schematics of the thalamocortical circuit and 138 relevant cell-types. (c) Spike rate aligned to transitions from non-whisking to whisking. (d) 139 Average spike rate as a function of whisking amplitude. (e) Average population response 140 aligned to touch. Data and figures corresponding to previously reported datasets [34, 35]. 141 142 VPM neurons have significant spike rates (mean, 5.1 Hz), even in the absence of whisker 143 movement and touch (Table 1; Fig. 1c, d; see S1 Fig, S2 Fig. for representative examples) [35]. 144 VPM neurons were also highly sensitive to whisking (S1 Fig. B,C). Spike rates increased after 145 whisking onset (average, 3-fold; Fig. 1c, d;). VPM neurons also responded to active touch with 146 a brief increase in spike rate (0.6 spikes per touch) (Fig. 1e). The peri-stimulus time histogram 147 (PSTH) aligned to touch onset shows a sharp peak in spike rate, with short latency after touch 148 (3.1±0.6 ms) and brief duration (2.9±1.7 ms) (S1 Fig. i,n,s). Exafferent touch signals and 149 reafferent whisking signals were multiplexed in individual VPM neurons (S1 Fig.). 150 The modulation of L4 FS neurons to touch and whisker movement were similar to VPM 151 neurons (Fig. 1c-e; S1 Fig., S2 Fig.). L4 FS neurons increased their spike rate after onset of 152 whisking (average, 3-fold). L4 FS neurons responded reliably to touch. The response onset had 153 slightly longer latency (5-15 ms) and longer duration (2-15 ms) than for VPM neurons. The 154 increased latency is expected from the propagation time delay between VPM and L4 (2 ms) 155 [35] [46]. Gutnisky et al 11/28/2016 9 156 The responses of L4 excitatory neurons differed profoundly from those of VPM and L4 FS 157 neurons [34] (Fig. 1 c-e, S2 Fig.). The baseline spike rate of L4 excitatory neurons was much 158 lower and did not increase significantly after onset of whisking. The response after touch onset 159 occurred with longer latency and longer duration than for the VPM neurons, and similar to L4 160 fast-spiking neurons (Fig. 1e). 161 The transformation performed by L4 circuits can be summarized by three main findings (Fig. 162 1c-e; Table 1). First, baseline spike rates are high in VPM and L4 FS neurons, and low in L4 163 excitatory neurons. Second, VPM and L4 FS neurons elevate their spike rates further during 164 periods of whisker movement, whereas L4 excitatory neuron spike rates remain low. The spike 165 rates of VPM and L4 FS neurons during periods of whisking are more than one order of 166 magnitude larger than those of L4 excitatory neurons. Third, all three neuron types respond to 167 touch reliably (Table 1; Fig. 1e). The L4 circuit therefore performs a behaviorally relevant 168 computation by propagating information related to external stimuli and suppressing responses 169 to whisker movement, a predictable stimulus. 170 Area Figure N Spikes/touch Non-whisking Whisking - - - spks spks/s spks/s VPM 4 29 0.6±0.5 5.1±5.5 13.5±12.7 L4 FS 4 18 1.9±1.0 9.3±8.5 21.1±15.8 L4e 4 46 0.3±0.4 0.4±0.6 0.6±0.9 Gutnisky et al 11/28/2016 10 171 Table 1. Summary experimental results for VPM, L4 excitatory and L4 FS neurons. Data are 172 taken from [35]. The increase of the average spike rate of L4 excitatory neurons from non- 173 whisking to whisking conditions was found to be insignificant. Values are reported as mean ± 174 standard deviation. 175 176 The dynamics of FS neurons during whisker movement suggest that these neurons suppress 177 whisker movement-related activity in L4 excitatory cells. Consistent with this view, 178 optogenetic reduction of activity in a subset of L4 FS neurons expressing the light-gated, 179 inhibitory chloride pump eNphHR3.0 (L4I-hr+) unmasks movement-related activity in L4 180 excitatory neurons [35] (Fig. 2a,b). Photostimulation of L4I-hr+ neurons decreases the activity 181 in L4 FS neurons on average and increases the activity in L4 excitatory neurons. The 182 suppression of L4I-hr+ neurons causes an increase in response in L4 excitatory neurons during 183 whisker movements (Fig. 2c) and an overall increase in the number of spikes per touch (Fig. 184 2d). This result implies that suppression of whisking response in L4 excitatory neurons 185 involves inhibition from FS neurons. 186 187 Fig. 2. Summary of effect of photoinhibition of L4 FS neurons on L4 neurons. (a) 188 Schematic of the hr+ experiment. A subset of L4 FS neurons do not express hr (hr- neurons). 189 (b) Photostimulation of hr+ neurons decreases the activity in L4 FS neurons and increases the 190 activity in L4 excitatory neurons. (c) Response of L4 excitatory neurons during whisker 191 movements. (d) Response to touch of L4 excitatory neurons. This figure is adapted from Yu et 192 al (2016). Gutnisky et al 11/28/2016 11 193 Membrane potential measurements provide additional mechanistic clues [35]. L4 excitatory 194 neurons depolarize substantially (6 mV) and briefly after touch. Touch-related inhibitory input 195 (from FS neurons) to L4 excitatory neurons is delayed by approximately 0.5 ms with respect to 196 excitatory input. It has been proposed that this short ‘window of opportunity’ [2, 33] allows L4 197 excitatory neurons to spike after touch, typically with one spike [34], before inhibition 198 suppresses L4 excitatory activity. During whisker movement excitation is matched by 199 inhibitory input, keeping the L4 excitatory neuron membrane potential well below spike 200 threshold. Suppression of self-movement signals is therefore implemented by inhibition within 201 L4. 202 Yu et al. (2016) also showed that activating halorhodopsin in L4I-hr+ neurons during whisker 203 movement suppresses activity in these neurons. This result poses a new question. The major 204 cellular effect of activation of halorhodopsin is hyperpolarization [47, 48]. Theoretical 205 investigations of cortical circuits, consisting of recurrently connected excitatory neurons 206 stabilized by inhibitory neurons, have investigated the effects of hyperpolarization of inhibitory 207 neurons [39]. Over a large range of conditions, hyperpolarization of inhibitory neurons leads to 208 increased spike rates in both excitatory and also inhibitory neurons. This "paradoxical effect" 209 is common to multiple network regimes, including networks with strong synaptic conductances 210 that fire in an asynchronous manner [38] and networks with moderate synaptic conductances, if 211 the excitatory-to-excitatory synaptic coupling gEE is sufficiently strong [39]. Does the L4 212 circuit lie outside of the modeled parameter regimes, or can other factors explain the lack of 213 paradoxical effect? 214 A computational model of tactile suppression in L4 Gutnisky et al 11/28/2016 12 215 To explain the dynamical response of the L4 circuits to thalamic input, and to understand the 216 roles of feedforward and recurrent connections, we constructed and analyzed a detailed 217 computational model of the L4 circuit. L4 excitatory and L4 FS neurons make connections 218 within type and across types with high connection probability (Fig. 3a; Table 2). 219 thalamocortical circuit of the rodent whisker system is one of the most extensively studied 220 mammalian circuits. In vitro and in vivo studies have measured many fundamental parameters 221 that are required to construct a realistic computational model of the thalamocortical circuit. Our 222 model circuit consists of 1600 L4 excitatory (E) neurons and 150 L4 FS (I) neurons [26] 223 receiving input from 200 VPM (T) neurons (Fig. 3a). When referring to modeling results we 224 denote VPM neurons as T, L4 excitatory neurons as L4E, and GABAergic interneurons as L4I 225 (Fig. 3b, c). Since we know little about non-FS GABAergic interneurons during behavior we 226 model only one inhibitory neuronal population. The 227 228 Fig. 3. Neural network model of L4. (a) Diagram of the recurrent model of L4 network. (b) 229 Spike shape of VPM (schematic), L4E and L4I neurons. (c) Temporal dynamics of individual 230 EPSPs for the different synaptic connections (T = VPM; I = L4 FS; E = L4E). The convention 231 is that that the first letter corresponds to the post-synaptic neuron and the second letter to the 232 presynaptic neuron. (d) Thalamic generating function FT (Eq. 1). The panels on the right show 233 the same figure in a magnified scale. 234 235 236 Gutnisky et al 11/28/2016 13 Populations Synaptic delay Kαβ gαβ Vextr Kαβ Vextr 2 Receptor (ms) mS/cm (mV) (mV) experimental experimental values values a ET AMPA 1.0 50 0.15 1.1 100 0.5a, 2.4±2b (response to first spike in a train), ~1b (spike train ~10 Hz). c IT AMPA 1.0 75 0.2 1.03 150 4.1±3b (response to first spike in a train), ~1b (spike train ~10 Hz), 3c EE AMPA 1.0 200 0.2 0.73 400e 1.1±1.1b, 0.52d, 1.6±1.6e IE AMPA 1.0 400 0.6 1.33 800b 2.2±2.2b EI GABAA 0.85 25 0.7 -1.92 50b -1.1b, -1.0f II GABAA 0.5 25 0.55 -1.28 50d -1.8f 237 238 Table 2. Network parameters in the reference parameter set: delay - synaptic delay , Kαβ - 239 average number of presynaptic inputs , gαβ - synaptic conductance, Vextr - the extremal value 240 of the unitary synaptic membrane potential change. The corresponding experimental values for 241 Kαβ and Vextr are written in the two right columns. Those values are taken from the following 242 references: a - [46], b - [49], c - [32], d- [26], e – [50], f - [25]. 243 244 Cortical neurons were simulated using a conductance-based model with a single compartment 245 per neuron [51, 52]. Synaptic conductances are denoted as gαβ, where β is the presynaptic 246 population and α is the postsynaptic population (Fig. 3a). Model parameters, including 247 numbers of neurons, unitary synaptic conductance, connection probability, resting potential, 248 and membrane time constants are based on neurophysiological measurements in brain slices 249 [25, 26, 49, 53] with small adjustments (see Methods). VPM provides the only external input Gutnisky et al 11/28/2016 14 250 to the model circuit, with connectivity estimates based on in vivo and brain slice measurements 251 [32, 46]. Spike trains of T neurons were modeled as independent inhomogeneous Poisson 252 processes with a generating function FT representing thalamic activity during quiescence, 253 whisking or whisking and touch (Fig. 3d; see Methods). 254 2 t C FT (t ) AT 1 BT sin T (t n w tc ) (n w tc c t ) w c 255 where AT is the spike rate averaged over a whisker movement cycle, BT is the modulation depth, 256 and CT is the number of spikes per touch, τw is the whisking period, tc is the time of touch onset 257 within a whisking cycle, n is the cycle number, and Θ is the Heaviside function. During 258 whisker movements AT increases above a baseline, with sinusoidal modulation phase-locked to 259 a single preferred phase ϕ. Touch is represented by adding a rectangular function at touch onset 260 t=tc (with respect to the whisking cycle), stretched over τc = 3 ms with an integral of CT =0.6 261 spikes per touch (Eq. 1). The population-average spike rate of neurons within a population over 262 whisking cycles is denoted by να (α=T,E,I), and the population-average spiking response to 263 touch is denoted by Rα. Note that without touch, νT = AT. A ‘reference parameter set’ is 264 described in Methods and used unless otherwise stated. 265 Response of the model circuit to whisker movement and touch 266 We start by simulating the model for whisking without touch (thalamic spikes per touch, CT=0). 267 L4E spike rates were low (νE < 1 Hz) compared to those of L4I neurons (Fig. 4a, b). The 268 average rates νE and νI depend on AT (the population- and time-average thalamic spike rate 269 during whisking only) but are almost independent of the modulation depth BT (Eq. 1). Except 270 near threshold, νI was proportional to AT, inconsistent with theoretical results for large and (1) Gutnisky et al 11/28/2016 15 271 sparse neuronal networks with strong synapses, in which strong excitation is compensated by 272 inhibition (balanced networks) [36]. The L4E spike rate νE increased linearly with AT, but with 273 a much smaller slope compared to νI. 274 275 Fig. 4. A neural network model of L4 explains suppression of whisker movement signals 276 in L4 excitatory neurons. The colors black, grey and red denote T, L4E and L4I neuronal 277 populations respectively. (a) Example L4E and L4I membrane potential during simulated 278 whisking (green). (b) The population- and time average spike rates νE and νI of the L4E and 279 L4I neurons respectively as function of the thalamic input AT in the absence of touch. L4I 280 neurons follow linearly the thalamic input while L4E neurons increase only weakly with AT 281 beyond firing threshold. Inset, zoom in. (c) Membrane potential for an example neuron during 282 whisking and touch (black dots). (d) Population PSTH aligned to touch onset. Inset, zoom in. 283 284 The linearity of the νI-AT curve fit empirical observations (Table 1, Figure 1c; spike rates of 285 both VPM and L4 FS neurons more than double in the transition from non-whisking to 286 whisking). The average spike rate of L4 excitatory neurons, νE, is low [35]. During transition 287 from non-whisking to whisking, the average spike rate of L4 excitatory neurons barely changes 288 (from 0.4 Hz to 0.6 Hz; Table 1). The spike rates of L4E neurons in our model show similar 289 dynamics. 290 L4E neurons fire at most one spike each after touch, whereas L4I neurons fire more than one 291 EI (up to two) spikes per touch (Fig. 4c, d). The synaptic delay, delay , is a critical factor in 292 EI determining the strength of the touch response (Fig. 5a). For delay 0 , the average normalized Gutnisky et al 11/28/2016 16 293 response of L4E neurons after touch, RE, is small (RE=0.01 spikes/touch), and L4I neurons fire 294 RI=0.64 spikes/touch. Strong inhibition overwhelms excitation before L4E neurons have a 295 chance to spike. The delay between the appearance of excitation and feed-forward inhibition in 296 the L4E neuron has been termed ‘window of opportunity’ [2, 33, 54-56]. This ‘window of 297 EI opportunity’ increases with delay , but is not equal to it because it is affected by the durations of 298 thalamocortical synaptic process and the time needed for the L4I to fire in response to the brief 299 EI and strong thalamic input. For a more realistic value, delay 0.85ms , L4E and L4I fire 0.33 300 and 1.3 spikes/touch respectively, similar to experimental measurements (Table 1). 301 302 Fig. 5. Touch response in function of synaptic delay, AMPA receptors’ time constant and 303 parameters defining thalamic input. In each panel, Response to touch in L4E and L4I 304 neurons (RE and RI, in grey and red respectively) are plotted, as well as the thalamic response 305 in black. Spikes per touch were counted up to 25 ms after touch onset, and baseline computed 306 by counting spikes 25 ms before touch is subtracted. Responses to touch are plotted as 307 EI functions of (a) I-to-E synaptic delay delay , (b) the AMPA receptor time constant tAMPA, (c) the 308 thalamic response to touch, CT, and (d) the thalamic spike rate AT during whisker movements 309 without touch. 310 311 The response of the network to touch depends on the time-course of synaptic conductances. 312 Without touch signals, L4E neurons are inhibited by L4I neurons and spike at low rates 313 (Fig. 4a, b). Touch produces strong, brief and synchronous thalamic excitation (Fig. 1e), which 314 depolarizes L4E neurons and enables them to fire before inhibition terminates the response. Gutnisky et al 11/28/2016 17 315 This mechanism demands that excitatory synaptic conductance changes at thalamocortical 316 synapses are brief. Touch responses of L4E and L4I neurons decrease with tAMPA, the decay 317 time of AMPA-mediated EPSCs (Fig. 5b). Substantial touch responses in L4E neurons require 318 tAMPA < 2-3 ms, consistent with the brief excitatory conductances measured at thalamocortical 319 synapses [53]. 320 Touch responses of L4E and L4I neurons increased with the strength of the thalamic touch 321 signal (Fig. 5c), consistent with graded responses to touch strength measured in L4E neurons 322 [34]. The response saturates at one and two spikes/touch for L4E and L4I respectively; this is 323 in part due to the intrinsic properties of these neurons, which preclude them from firing more 324 spikes in response to brief thalamic input. 325 We have shown that a L4 network with parameters similar to experimentally determined 326 values can replicate major experimental findings. L4E excitatory neurons exhibit low baseline 327 activity, low activity in response whisker movement, and significant response to touch. 328 However, during behavior the responses of L4 excitatory neurons is not all-or-none: L4 329 excitatory neurons are tuned to multiple sensory features including touch direction, intercontact 330 interval, strength of touch, and likely other factors [34]. Our simulation results indicate that 331 L4E and L4I neurons respond to touch in a graded manner and robustly respond to touch at 332 different whisking amplitudes. Specifically, our simulations show that RE and RI decrease with 333 AT (Fig. 5d) for all parameter values consistent with the data. Mechanistically, increasing 334 thalamic input causes larger inhibition in the cortical circuit that decreases touch response. 335 Future experiments could test this model prediction. 336 Contributions of intracortical connections in tactile suppression Gutnisky et al 11/28/2016 18 337 The model allows us to explore how specific synaptic connections within L4 contribute to fine- 338 tuning L4 function. These connections currently cannot be specifically manipulated, making 339 them difficult to evaluate experimentally. Recurrent excitation (gEE) can only be tuned over a 340 limited range before runaway excitation is triggered (for about gEE~0.4 mS/cm2), even during 341 baseline or whisking (Fig. 6a-b). The conductance gEE does not modify the slope of the νE-AT 342 curve (Fig. 6b), but shifts this curve upward. gEE amplifies touch responses RE and RI (Fig. 6c) 343 [57], mostly by increasing the duration of touch responses (Fig. 6d-i). 344 345 Fig. 6. Effects of varying intracortical recurrent excitatory conductances gEE on the 346 function of L4E neurons. (a) The circuit with changing gEE emphasized in green. (b) RT, RE 347 and RI as functions of gEE. Other parameters: AT = 14 spikes/s, CT = 0.6). (c) νE vs. AT during 348 whisker movements only, for 11 values of gEE from 0 (light green) to 0.4 mS/cm2, that is twice 349 the reference parameter value (dark green). Recurrent excitation gEE increases νE while not 350 affecting the slope of the νE-AT curve far from spiking threshold substantially. (d) PSTH 351 aligned to touch onset for L4E and without recurrent excitation (gEE=0 mS/cm2). (e) Same as C 352 for L4I. (f-g) Same as c-d for gEE=0.2 mS/cm2. (h-i): Same as c-d for gEE=0.35 mS/cm2. 353 Beyond ~gEE=0.4 mS/cm2 the network exhibits runaway excitation. 354 355 The dynamics of L4E neurons are shaped in subtle ways by L4 inhibition. Figure 7 shows how 356 parameters involving inhibition (gII, gEI, gIE) affect circuit responses, while holding the other 357 parameters at their reference values. Intracortical excitation of inhibition (gIE) is necessary to 358 prevent runaway excitation, and thus keeps the spike rates of neurons in the network moderate Gutnisky et al 11/28/2016 19 359 (Fig. 7a,b). Reducing gIE shifts the νE-AT curve during whisking upward, without modifying its 360 slope. 361 decreasing the slope of the linear section of this sigmoid (Fig. 7c). For touch responses, gIE shift the touch-response RE-CT curve to the right while 362 363 Fig. 7. Inhibition in L4 controls the response to L4E neurons. (a-c) Changing gIE from 0 to 364 1.2 mS/cm2 (d-f) Changing gEI from 0 to 1.4 mS/cm2. (g-i) Changing gII from 0 to 1.1 mS/cm2. 365 In the top panels, the synaptic connection that its strength is varied is plotted in green. In the 366 middle and bottom panels, curve are plotted for 11 values of the relevant g from 0 (light green) 367 to its maximal value, that is twice the reference parameter value (dark green). (b, e, h) The 368 response of L4E neurons, νE, to slowly varying thalamic input, which correlates with the 369 amplitude of whisking. (c, f, i) The response of L4E neurons to spikes, RE, associated with 370 touch. 371 372 Inhibition to L4E neurons (gEI) is also required to prevent runaway excitation (Fig. 7d,e). For 373 low values of gEI (but above values for runaway excitation) and large values of gII (Fig. 7g,h), 374 νE scales linearly with thalamic input AT. This linear scaling is known from balanced networks 375 with strong synaptic coupling [36, 38]. The situation is different for sufficiently strong 376 inhibition (relatively high gEI, moderate gII), where the response of L4E neurons to slowly- 377 varying thalamic input during whisking, νE, is independent, and even decreasing, with AT (Fig. 378 7e, h). For small enough gII, L4E are quiescent except of near firing threshold. In addition, low 379 gET and large gIT values are critical to keep νE low for all whisker movement amplitudes (Fig. 380 8a,b,d,e). Furthermore, these low values νE are nearly independent of AT. This regime is Gutnisky et al 11/28/2016 20 381 consistent with experimental data, where L4 FS neurons increase their spike rate with whisker 382 movement, whereas L4 excitatory neuron spike rates remain low (Fig. 1d, Table 1). 383 384 Fig. 8. Effect of varying thalamocortical conductances on the function of L4E neurons. 385 Symblols and lines are as in Figure 7. (a) Changing gET.. (b) νE vs. AT during whisker 386 movements only. (c) RE vs. CT. (d) Changing gIT. (e-f) Same as b-c. 387 388 Increasing values of gEI and gII shifts the touch-response RE-CT curve to the right and to the left 389 respectively (Fig. 7f, i). In addition, gEI, but not gII, decreases the slope of the linear section of 390 this sigmoid (Fig. 7f). Similarly, gET and gIT shift the RE-CT curve to the left and to the right 391 respectively (Fig. 8c, f). 392 The effects of halorhodopsin activation in L4 FS neurons 393 Yu et al. (2016, Fig. 8, supplementary Fig. 13) expressed halorhodopsin in FS neurons in L4 394 and explored how the responses of excitatory and FS neurons in L4 to under baseline 395 conditions, whisking and touch vary under halorhodopsin activation. L4 excitatory neurons 396 increase their spike rates and L4 FS neurons decrease their spike rates, both during baseline 397 and whisking conditions. Similarly, L4 excitatory neurons increase their spiking responses to 398 touch and L4 FS neurons decrease their responses. 399 The halorhodopsin expression levels in FS neurons likely varied widely across individual 400 neurons [58]. We therefore divided the L4I neurons in the model into hr expressing neurons Gutnisky et al 11/28/2016 21 401 and non-expressing neurons (hr+ and hr- respectively). The fraction of hr+ neurons among L4I 402 neurons is denoted by fhalo. 403 404 Manipulation of the components of neural networks can cause complex and counterintuitive 405 change in network dynamics (‘paradoxical’ response; [38-40, 59]: injecting negative current to 406 all inhibitory neurons in a network, that includes spiking excitatory neurons, increases the 407 average spike rates in inhibitory neurons). If the synaptic conductance strengths are moderate 408 and the excitatory-to-excitatory synaptic conductance is above a certain level, injecting 409 negative current to the inhibitory neurons causes the spike rates of excitatory neurons νE to 410 increase, and as a result the network dynamics causes the spike rates of inhibitory neurons νI to 411 increase as well [39, 40, 59]. Alternatively, if synaptic conductances are strong and the 412 excitatory population is active, the condition that the activity of excitatory neurons should be 413 moderate (non-zero and not epileptic) causes νI to increase under negative current injection to 414 inhibitory interneurons [38]. The major effect of halorodopsin is to hyperpolarize FS neurons 415 (see Methods). Therefore, if all inhibitory neurons in the model are assumed to express 416 halorodopsin equally (fhalo=1), halorhodopsin activation will cause both L4E and L4I neurons 417 to increase their spike rates. 418 419 We simulated the response of L4E and L4I neurons to whisking for fhalo=1 without and with 420 halorhodopsin activation, and replicated the 'paradoxical effect': the average responses of L4E 421 and L4I neurons to whisking, νE and νI, increase with simulated light activation. The simulation 422 result for L4I neurons is in contrast to experimental observations showing that most L4 FS 423 neurons increase their spike rates. This discrepancy is resolved if we assume fhalo =0.5. For this Gutnisky et al 11/28/2016 22 424 value, spike rates during whisking increase, on average, for L4E neurons (Fig. 9a), decrease 425 for almost all L4I-hr+ neurons (Fig. 9b), and increase for inhibitory neurons that do not express 426 halorhodopsin (L4I-hr-). 427 428 Fig. 9. Simulated light activation of halorhodopsin expressed in L4I-Hr+ neurons. 429 Simulations with fhalo=0.5 reveals a reduction in the whisking suppression and an enhancement 430 of touch responses by L4E neurons. (a) Halorhodopsin activation in L4I-Hr+ causes an average 431 increase in response of L4E during whisking and no touch, with a wide distribution of 432 halorhodopsin–induced modifications. (b) Most L4I-Hr+ neurons reduce their activity during 433 whisking while L4I-Hr- neurons increase it. (c) Increase in the touch responses in L4E neurons 434 during suppression of L4I-Hr+. (d) Increase in the touch responses in L4I neurons. (e,f) 435 Population PSTH of L4E (e) and L4I (f) neurons with and without L4I-Hr+ activity suppression. 436 (g) Reduction of L4I-Hr+ activity diminishes the whisking suppression effect in L4E neurons. 437 Black line: T neurons; solid grey line: L4E neurons without halorhodopsin activation; dashed 438 grey line: L4E neurons during halorhodopsin activation. (h) Reduction of L4I-Hr+ activity 439 diminishes the whisking response in L4I-Hr+ neurons. Solid red line: L4I-Hr+ neurons without 440 halorhodopsin activation; dashed red line: L4I-Hr+ neurons during halorhodopsin activation. 441 Solid blue line: L4I-Hr- neurons without halorhodopsin activation; dashed blue line: L4I-Hr- 442 neurons during halorhodopsin activation. 443 444 445 The spike responses of L4E neurons to touch increase moderately for fhalo =0.5, especially for 446 neurons with low baseline response (Fig. 9c, e), whereas the spike response of and L4I-hr- Gutnisky et al 11/28/2016 23 447 inhibitory neurons remains about the same and that of L4I-hr+ increases somewhat (Fig. 9d, f). 448 These dynamics stem from the fact that the initial response to touch, before feedforward 449 inhibition hyperpolarizes the neuron, is determined mainly by feedforward excitation. While 450 halorhodopsin activation hyperpolarizes inhibitory neurons and increase the difference 451 between spiking threshold and their membrane potentials before touch onset, the driving force 452 of excitation (the difference between the reversal potential of APMA-mediated excitation and 453 their membrane potential) increases as well and enhances excitation. 454 455 The reduction in response in L4I-hr+ neurons partially disrupts whisker movement suppression 456 in L4E neurons, and increases the slope of the νE-AT curve (Fig. 9g). During photostimulation 457 of L4I-hr+ neurons, their response to increasing whisker movements becomes shallower (Fig. 458 9h). This change in response of L4-hr+ neurons in turns causes a steeper response to whisker 459 movements in L4E neurons (Fig. 9g). 460 461 Finally we explored how the previous results depend on fhalo. The mean touch responses for all 462 neuronal populations is non-monotonic: it increases when fhalo increase from zero, and then 463 decreases for higher values (Fig. 10a). For our parameter set, simulated halorhodopsin 464 activation increases touch response (in comparison with no activation) for fhalo < 0.78. During 465 whisker movements, the spike rates of all neuronal populations increases with fhalo, and the 466 average spike rate of L4I-hr+ neurons during simulated halorhodopsin activation is lower than 467 that with no activation for fhalo < 0.74 (Fig. 10b). Interestingly, L4-hr+ neurons reduce their 468 spiking response to touch in response to halorhodopsin activation for large values of fhalo, but 469 reduce their spike rates in response to whisking for small fhalo values. Gutnisky et al 11/28/2016 24 470 Fig. 10. Effects of light activation of halorhodopsin expressed in L4I-Hr+ neurons. Effects 471 of halorhodopsin light activation on touch response and whisking response for the L4E (grey), 472 L4I-Hr+ (red) and L4I-Hr- (blue) neuronal populations are plotted as functions of fhalo, the 473 fraction of L4I-Hr+ neurons among all L4I neurons. (a) Spikes per touch, R, for L4 neuronal 474 populations as functions of fhalo. (b) Average spike rates, ν, during whisking for L4 neurons as 475 functions of fhalo. 476 477 478 Discussion 479 We computationally dissected how information is transformed between thalamus and barrel 480 cortex at the level of defined neural circuits [34, 35] (Fig. 1). Our model defines the 481 mechanisms that enable the L4 circuit to transmit touch-related information and suppress self- 482 motion signals in L4. Here we summarize our main modeling results and how they explain 483 major experimental observations. 484 Transmission of touch signals while suppressing whisking signals 485 We compared our model to recordings from VPM projection neurons (Fig. 1; S1 Fig., S2 Fig.), 486 L4 FS neurons, and L4 excitatory neurons [35] during performance of an object localization 487 task, while whisker movements and touches were tracked with millisecond time scale precision 488 [34, 42, 45, 60]. VPM and L4 FS neurons respond to whisker movement and touch, whereas 489 L4 excitatory neurons responded almost exclusively to touch. During whisking, excitation to 490 L4 excitatory neurons from VPM is only slowly modulated in time and is matched by 491 feedforward and feedback inhibition from L4 FS neurons, which cancels self-movement Gutnisky et al 11/28/2016 25 492 signals. In contrast to the self-movement input, touch-related inputs are brief and synchronous 493 (Fig. 1e). Our model establishes the conditions for a ‘window of opportunity’, in which L4 494 excitatory neurons can fire before inhibition catches up [33, 55, 57, 61, 62]. The I-to-E 495 EI synaptic delay delay must be sufficiently large (say, ~1 ms) and the AMPA-mediated synaptic 496 conductances should be brief (Fig. 6a,b). The brief duration of this window diminishes the 497 chance of L4 excitatory neurons to fire multiple spikes upon touch, producing low trial-to-trial 498 variability in spike count after touch [34, 63]. 499 Spike rates during baseline and whisking 500 During baseline and whisking, L4 FS neurons spike on average at tens of Hz, while L4 501 excitatory neurons spike on average below 1Hz. The average spike rates of VPM and L4 FS 502 excitatory neurons, νT and νI, more than double after whisking onset, while the average spike 503 rates of L4 excitatory neurons remains below 1 Hz. Low spike rates of L4E neurons require 504 particular combinations of synaptic parameters: gIE,, gEI and gIT need to be strong, and gEE,, gII 505 and gET should be weak (Figs. 6c,7,8). If gET is too small or gET is too large, L4E neurons will 506 be quiescent. Similarly, L4E neurons will not spike if gII is too small (Fig. 7h). For moderate 507 gII, νE can remain about constant with AT. 508 Touch responses 509 The response of L4E neurons to touch, RE, increases with CT (the thalamic response to touch) 510 in a sigmoid manner, because the intrinsic properties of L4E neurons in the model impose a 511 refractory period that in general precludes rapid firing of more than one spike. Synaptic 512 parameters that increase νE during whisking shift the sigmoid function leftward, and those that Gutnisky et al 11/28/2016 26 513 decrease νE shift that function rightward (Figs. 7,8). In our model, the inhibitory neuronal 514 population spikes at significant rates (20-40 Hz) even at baseline. This allows the number of 515 spikes/touch RE to be smaller than one and vary gradually with input strength (Fig. 7f). As a 516 result, stronger thalamic responses to touch, for example as a result of stronger touch, 517 generates proportionally stronger L4E responses, consistent with experimental results (Fig. 1e) 518 [34]. 519 Effects of photostimulation of L4-hr+ neurons 520 In response to halorhodopsin activation, L4 excitatory neurons increase their average spike 521 rates and L4-hr+ FS neurons reduce their spike rates [35]. The model does not replicate this 522 behavior if halorhodopsin is "expressed" in all inhibitory neurons, because νI increases with 523 simulated photostimulation. This discrepancy is resolved if a sufficient number of L4 I neurons 524 lack halorhodopsin (or express halorhodopsin at very low levels) (Figs. 9, 10b). With 525 photostimulation, L4-hr+ I neurons decrease their spike rates if their fraction among I neurons, 526 fhalo, is below a certain value (0.74 in Fig. 10a), while L4-hr- FS neurons increase their spike 527 rates. Future work is needed to determine whether including more populations of inhibitory 528 interneurons, for example somatostatin-positive neurons [64] will generate parameter regimes 529 in which L4-hr+ neuron decrease their responses to both whisking and touch upon light 530 activation. 531 Modeling approach 532 The emergence of powerful computing resources has enabled simulations of entire cortical 533 columns of diverse neurons with realistic morphologies and biophysically plausible membrane Gutnisky et al 11/28/2016 27 534 conductances [65, 66]. Because of the profusion of parameters, detailed models are difficult to 535 analyze to extract the underlying principles. So far these detailed models have failed to explain 536 any neural computation [67]. Here we took a different approach. We implemented a 537 computational model based on hard won numbers for connection probability, connection 538 strength, neuron numbers and basic cellular parameters for the whisker thalamocortical circuit. 539 The neurons themselves were generic single compartment models. Given the reduced nature of 540 the model, it is possible to analyze the model in detail. This analysis yields testable predictions. 541 For example, the dependence of average spike rates of neuronal populations on the average 542 spike rates of thalamic neurons (Fig. 4b). These predictions could be tested in experiments 543 where stimuli, applied for example with a magnetic stimulator, are applied during whisker 544 movement at different amplitudes. 545 The main characteristics of the thalamocortical circuit are: (i) Strong external inputs to fast- 546 spiking inhibitory neurons and excitatory neurons. (ii) Strong inhibition within L4, 547 implementing feedforward and lateral inhibition. (iii) Recurrent excitation. iv) A brief but 548 nonzero delay between intracortical inhibitory and external excitatory input. Parameters based 549 from neurophysiological measurements produce a model circuit with behavior that is in 550 qualitative agreement with in vivo measurements over a large range of conditions. However, it 551 was necessary to adjust the model parameters slightly (i.e. within a factor of 2-3; Methods) to 552 best match experimental data (reference parameter set, Methods). Experimental inaccuracy, 553 sampling errors, and differences across biological conditions (such as in vitro brain slices 554 versus behaving brain) likely preclude more accurate estimation of synaptic parameters. 555 Changing even a single parameter could lead to quantitative, and sometimes even qualitative 556 changes in behavior (Fig. 6-8), as has been noted in other contexts [68]. This means that the Gutnisky et al 11/28/2016 28 557 detailed parameters matter. For example, if gII is too strong then L4I neurons do not respond 558 sufficiently briskly to reduce whisking signals in L4E neurons (Fig. 7g, h); on the other hand, 559 if gII is too weak then inhibition shuts down L4E neurons. Similarly, gEE > 0 is required to 560 amplify touch signals, but elevating gEE by a factor of two beyond the optimal level causes run- 561 away excitation (Fig. 6b). Moreover, the kinetics matter, such as the durations of the I-to-E 562 EI synaptic delay delay (Fig. 5a), axonal delays, the time-courses of synaptic conductances (Fig. 563 5b), and intrinsic neuronal properties. 564 Comparison with previous L4 models 565 Rate models have been used to study cortical responses to fast-rising stimuli in the whisker 566 system [8, 54] and elsewhere [64]. These previous models of L4 aimed to account for fast 567 cortical responses to passive whisker deflections [8], but rate models cannot reliably describe 568 brief responses to rapidly-varying stimuli [69]. Moreover, in important aspects these models 569 have opposite behavior to our experimental and modeling results and to the known anatomy 570 and physiology of L4 circuits. First, the modeled L4 neurons show activity in the absence of 571 input, in contradiction to recent measurements [35, 70]. Second, L4 neurons exhibit strong and 572 EI brief response to touch even for delay 0 , in contrast to our model. 573 The propagation of synchronous and brief activity in cortical circuits has been investigated [2, 574 56, 57, 71, 72]. In these models, excitatory inputs produce strong inhibition in local circuits, 575 which is slightly delayed with respect to the excitation. In L4 of the barrel cortex, the 576 convergence of thalamocortical input onto L4 FS neurons is higher than for L4 excitatory 577 neurons (probabilities of connections are 0.75 vs 0.4) [32] and thalamic stimuli produce larger 578 synaptic potentials in L4 FS than L4 excitatory neurons [33]. These features allow propagation Gutnisky et al 11/28/2016 29 579 of brief synchronous activity across network layers. These models can respond to brief, strong 580 stimuli by brief responses of the L4 E and I neuronal populations. Our model goes beyond this 581 effect by showing how this brief touch response can be obtained together with large νI and 582 small νE in response to baseline and whisking. For example, simple feedforward models such 583 as [33] cannot exhibit small νE for a wide range of AT, corresponding to both baseline and 584 whisking. That model, however, explores the development of cortical responses of L4E and 585 L4I neurons over the long time scale of short-term synaptic plasticity. Conducting of such a 586 study in a model that includes recurrent connections and comparing its outcome to 587 experimental result in vivo remains to be carried out. 588 Models of networks of strongly-coupled neurons compensated by inhibition have been studied 589 extensively [36-38, 51, 73]. If inhibition on the excitatory neurons is not too large, the 590 population-average response of excitatory and inhibitory neuron (νE and νI) scales linearly with 591 the external (here, thalamic) input AT [37, 74]. The slope of the νE vs. AT increases with gII and 592 decreases with gIE.. If inhibition is too strong, νE =0 and νI scales linearly with AT. In our 593 network, the numbers of neurons and the strengths of synaptic conductances are not very large. 594 As a result, when νE is small (νE 595 of near firing threshold, namely the AT values for which cortical neurons start firing) increase 596 weakly with AT, be independent of AT, or even decrease with AT (Fig. 7e,h, Fig. 8b,e). Similar 597 behavior was obtained in rate models of V1 [75, 76]. This result mimics the empirical 598 observation that during transition from non-whisking to whisking, the spike rates of VPM 599 neurons more than double, the spike rates of L4I neurons increase proportionally, but the spike 600 rates of L4E neurons remains low, with no significant increase with thalamic input strength. 601 For larger νE (> 2 Hz), νE depends linearly on AT. Larger νE could be obtained with larger gII or ~ 1 Hz), similar to measured values [34, 35], νE can (except Gutnisky et al 11/28/2016 30 602 smaller gEI (Fig. 7b,e,h), or by suppressing a fraction of L4I neurons using halorhodopsin 603 activation (Fig. 9g). 604 Based on the linearity of strongly-coupled networks, one would expect that if responses of L4E 605 neurons to whisking is an order of magnitude smaller than that of VPM neurons, the same 606 proportions will maintain for the responses to touch, whereas experimentally the touch 607 responses of L4 excitatory and VPM neurons differ by only a factor of two. We explain this 608 EI effect by the fact that, with delay , inhibition lags excitation in response to brief and strong 609 EI thalamic input. Indeed, for delay 0 , the response to L4E neurons is significantly reduced (Fig. 610 5a). 611 612 Materials and Methods 613 Experimental methods 614 The experimental methods were described in Yu et al (2016). In brief, we performed chronic 615 multi-electrode silicon probe recordings from VPM and cell-attached recordings from L4 616 excitatory and L4 fast spiking neurons. To search for VPM neurons we lightly anesthetized the 617 mice (0.6-1.2 % isoflurane) and stimulated individual whiskers during extracellular recordings. 618 We mapped the principal whisker (PW) to assess whether the PW was one of the large 619 whiskers that can be tracked reliably during behavior. We stimulated several individual 620 whiskers around the PW with a piezoelectric stimulator at multiple frequencies (i.e. 5,10,20,40 621 Hz) and recorded the neural activity. Before and/or after each day of behavioral recording we 622 confirmed that neurons respond to stimulation of the PW (we re-positioned the electrode drive 623 daily). In addition, after placing the animal in the behavioral apparatus, we usually maintained Gutnisky et al 11/28/2016 31 624 the anesthesia for several minutes to check that neurons still responded to the PW (by manual 625 stimulation and/or by contacting the whiskers with the pole). 626 behavioral task shortly after anesthesia was withdrawn. 627 Statistical analyses. Data is presented as mean ± s.d. or mean ± s.e.m. as noted. Statistical 628 comparisons use the non-parametric, two-sided Wilcoxon rank sum test. All statistical analyses 629 were performed in Matlab (Mathworks). Animals performed the 630 631 Behavior, videography and recordings. In Yu et al (2016) we recorded from VPM neurons in 632 mice trained to localize an object using their whiskers [42]. In each trial, during a sample 633 epoch lasting a few seconds (1.1 - 2.5, mean 1.9 s), a pole appeared in one of multiple 634 locations on the right side of the head (Fig. 1a). Animals had to withhold licking when the pole 635 was presented in the most anterior location (typically out of reach of the whiskers), or lick to 636 obtain a reward when the pole was in the posterior locations. High-speed videography and 637 automated whisker tracking quantified whisker movement (azimuthal angle, ; whisking phase, 638 ), changes in curvature caused by the forces exerted by the pole on the whisker (change in 639 curvature, ∆)[34, 44, 77], and contact time, all with 1 millisecond temporal precision [43, 45]. 640 Mice touched the pole multiple times (mean number of touches, 8.9±4.2 before reporting 641 object location with licking (mean reaction time 416 ± 165 ms; mean ± s.d.). Recordings were 642 made in trained mice (71±7 % correct responses). To examine the relationship of VPM activity 643 and tactile behavior we aligned spikes of individual neurons recorded in VPM with high-speed 644 recordings of whisker movement and touch [34, 44, 45]. 645 646 Gutnisky et al 11/28/2016 32 647 Neuronal network model of a layer 4 barrel 648 Comprehensive neuroanatomical and neurophysiological data sets are revealing the 649 connectivity between defined cell types over multiple spatial scales [26, 78-82]. But links 650 between neural representations, computation and detailed anatomy are rarely achieved [6, 83]. 651 One challenge is a lack of knowledge about strengths and dynamics of synapses between 652 specific cell types during behavior. Most of such studies were carried out in slices, and there 653 are differences between intrinsic and synaptic properties measured in slices and in vivo and 654 between different in vivo states. Therefore, our strategy is to set parameter values (synaptic 655 conductances, connectivity) close to measured in vitro values. However, differences between 656 in vitro and in vivo conditions and experimental errors (e.g. estimates of unitary synaptic 657 strengths) currently preclude exact quantification of the synaptic and connectivity properties 658 (e.g., [46]). We define a set of parameter values, named "reference parameter set" [84], close to 659 measured values when known (Fig. 3), allowing for adjustment over a restricted range (mostly 660 within a factor of 2 relative to empirical values), such that the circuit displays dynamics similar 661 to experimentally-observed behavior. The reference parameter set specified below is used 662 unless otherwise stated. Then, we vary one or two parameters to explore the role of those 663 parameters on the system dynamics (e.g., Figs. 6-8). 664 Network architecture. We modeled one barrel in layer 4 (L4) of the vibrissa primary 665 somatosensory cortex (vS1). Our model circuit consists of L4 excitatory (E) neurons and L4 666 FS (I) neurons [26] receiving input from VPM (T) neurons (Fig. 3a). When referring to 667 modeling results we denote VPM neurons as T, L4 excitatory neurons as L4E, and GABAergic 668 interneurons as L4I (Fig. 3b, c). Non-FS interneurons have not been studied during behavior 669 and are therefore not included in the model. Gutnisky et al 11/28/2016 33 670 671 672 Dynamics of single cortical neurons. Intrinsic neuronal properties of L4 neurons have not been 673 fully characterized. Therefore, following previous work on cortical dynamics [51, 52], single 674 neurons are governed by a modified Wang-Buzsáki model [51, 52] with one compartment. The 675 membrane potential Vi , where α=E,I and i=1,…, Nα obeys dVi C I L, i I Na, i I Kdr,i I KZ,i I syn,i dt 676 (2) 677 678 where C is the neuron capacitance, 679 I L, i gL (Vi VL ) 680 (3) 681 is the leak current, the conductance of the leak current is gL=0.05 mS/cm2 and 0.1 mS/cm2 for 682 the excitatory and inhibitory neurons, respectively, and VL = -65 mV. The current I syn,i is the 683 3 total synaptic input into the neuron. The other ionic currents are: I Na g Na m h(V VNa ) where 684 m m (V) / ( m (V) m (V)) , I Kdr g Kdr n 4 (V VK ) , and I KZ gKZ z(V VK ) . The kinetics 685 of h and n are given by: 686 687 dh / dt h (V )(1 h) h (V )h , dn / dt n (V )(1 n) n (V )n (4) The functions α(V) and β(V), for V in mV, are: 688 689 m (V ) 0.1(V 30) /{1 exp[0.1(V 30)]} (5) 690 m (V ) 4exp[(V 55) /18] (6) Gutnisky et al 11/28/2016 34 691 h (V ) 0.7exp[(V 44) / 20)] (7) 692 h (V ) 10 / {1 exp[0.1(V 14)]} (8) 693 n (V ) 0.1(V 34) / {1 exp[0.1(V 34)]} (9) 694 n (V ) 1.25exp[(V 44) / 80] . 695 We set ϕ=0.2 to produce wider spikes than in [52]. The gating variable z of the adaptation 696 current IKZ is governed by the equation: (10) 697 698 dz / dt [ z (V ) z ] / z (11) 699 700 Where z (V) 1/{1 exp[0.7(V 30)]} and τz = 60 ms. The parameters of the model are: 701 gNa=100 mS/cm2, VNa= 55 mV, gKdr=40 mS/cm2, VK= -90 mV, C=1µF/cm2. Excitatory neurons 702 have gKZ=0.5 mS/cm2, whereas inhibitory neurons do not possess this current (gKZ=0). 703 704 Network architecture. The network consists of NE excitatory neurons and NI inhibitory neurons. 705 The L4 neurons receive input from NT thalamic relay neurons in one barreloid. For the 706 simulations, NE = 1600, NI = 150, and NT = 200 [26]. Each neuron, whether excitatory or 707 inhibitory, receives excitation and inhibition from the excitatory and inhibitory cortical 708 neuronal populations respectively, as well as excitation from VPM thalamic neurons. The 709 probability that a neuron from the βth, pre-synaptic population forms a synapse on a neuron 710 from the αth, post-synaptic population is K / N [85]. Therefore, a neuron from the αth 711 population receives, on average, Kαβ synaptic inputs from neurons in the βth population. We Gutnisky et al 11/28/2016 35 712 define the matrix Cij to be 1 if the jth neuron from the βth population projects to the ith 713 neuron from the αth population, and 0 otherwise. 714 715 Synaptic input. We consider only fast synapses corresponding to AMPA and GABA A receptors. 716 The total synaptic input a cell i from population α receives is the sum of the synaptic currents 717 from all the pre-synaptic populations: I syn, i 718 T,E,I Gsyn, i (t ) Vi Vsyn (12) 719 720 , The total conductance Gsyn,i for β = T,E,I is: 721 Gsyn, i (t ) syn,all K N g Cij s (t ) j (13) j 1 722 where the normalization constant τsyn,all is set to be 1 ms. The value s denotes a synaptic j 723 variable of the pre-synaptic neuron j from the βth population projecting to the αth population. 724 Those variables for AMPA and GABAA conductances are: 725 s j (t ) 1 tsyn exp (t t k k delay ) / tsyn (t tk delay ) (14) 726 The pre-synaptic neuron fires at times tk , k is the spike index, tsyn is the time constant of 727 synaptic decay, Θ is the Heaviside function, and delay is the time delay between the firing by a 728 spike from a presynaptic neuron from the βth population and the onset of the post-synaptic 729 conductance in the post-synaptic neuron from the αth population. Unless specified otherwise, 730 we use time constants tsyn = tAMPA = 2 ms for AMPA-mediated synapses and tsyn = tGABAA = 3 ms 731 for GABAA-mediated synapses. The reversal potentials are VAMPA=0 and VGABAA=-85 mV. Gutnisky et al 11/28/2016 36 732 The values of gαβ, Kαβ and delay are based on [25, 26, 32, 46, 49, 53] (Table 2). Compared to 733 literature values the Kαβ were reduced by a factor of 2 to counteract strong synchrony in the 734 cortical circuits. Alternatively this could have been achieved by modeling highly variable 735 synaptic strengths [86]. 736 The experimental evolution of the response of VPM and L4 neurons to consecutive touch 737 events has not reported yet. In this work, we do not address how the responses to whisking and 738 touch vary over the long time scale of short-term plasticity. For these reasons, and to keep the 739 model and its analysis simple, the synapses in the model do not have depression and 740 facilitation properties. 741 Spike rates of thalamic neurons. Spike trains of T neurons were modeled as statistically- 742 independent inhomogeneous Poisson processes. All T neuron share the generating function FT 743 mimicking thalamic activity during quiescence, whisking or whisking and touch (Fig. 3d): 744 2 t C FT (t ) AT 1 BT sin T (t n w tc ) (n w tc c t ) w c 745 The first term in Eq. 1 represents the baseline and whisking modulation contribution to the 746 instantaneous spike rate, and the second term represents spikes added by touch. The parameter 747 AT is the spike rate averaged over a whisker movement cycle without touch, BT is the 748 modulation depth, and CT is the average number of additional spikes per touch, τw is the 749 whisking period, tc is the time of touch onset within a whisking cycle, and n is the cycle 750 number. During whisker movements, AT increases above a baseline, with sinusoidal 751 modulation phase-locked to a single preferred phase ϕ (several neurons in VPM exhibit phase- 752 locking activity to the whisking cycle; Yu et al, 2016). Touch is represented by adding a (15) Gutnisky et al 11/28/2016 37 753 rectangular function at touch onset t=tc (with respect to the whisking cycle), stretched over 754 3 ms with an integral of CT =0.6 spikes per touch (Eq. 1). We use the parameters: BT = 0.25, τc 755 = 3 ms, T=100 ms, ϕ=π/2. The parameter tc is set to be 0.5T because thalamic neurons reach 756 their maximal spike rate during whisking at around maximal retraction (Yu et al., 2016). The 757 parameters AT and CT are 6 Hz and 0 during no-whisking states, 14 Hz and 0 during whisking 758 states and 14 Hz and 0.6 during whisking-and-touch states. For simplicity, we take the same 759 values of AT, BT and CT for all thalamic neurons. Because of the large convergence of thalamic 760 inputs into L4E and L4I neurons, our results are not modified significantly if we allow those 761 thalamic parameters to be taken from a uniform distribution with the same average values. 762 763 Population- and time-averaged quantities. We define να to be the population- and time- 764 averaged spike rates of all the neurons in the αth population over many whisking cycles. Touch 765 response of a neuron is defined to be the difference between the number of spikes fired by that 766 neuron during a time window of 25 ms after touch onset and the number of spikes fired by the 767 neuron 25 ms before touch onset. We define Rα to be the population- and time-average of the 768 spiking response to touch of all the neurons in the αth population over many touch events. 769 770 Simulations of halorhodopsin activation. We model activation of halorhodopsin, a light-gated 771 chloride pump, expressed in L4I FS neurons (Fig. 9,10). Halorhodopsin expression levels 772 likely varied widely across individual neurons [58]. We therefore assume that halorhodopsin is 773 expressed in only a fraction fhalo of the inhibitory neurons, 0 f halo 1. Activation by light of 774 the ith FS neurons that express halorhodopsin was modeled by two processes. First, as a 775 chloride pump, its activation has an effect of negative current injection to the neuron. The Gutnisky et al 11/28/2016 38 776 amplitude of this current in the ith neuron is I halo,i I halo,0 I halo,1 xi , where xi is taken randomly 777 from a uniform distribution between -1 and 1, Ihalo,0=-2 µA/cm2, and Ihalo,1=1 µA/cm2. As a 778 result, I halo,i varies between -3 µA/cm2 and -1 µA/cm2. Second, the GABAA reversal potential 779 is depolarized [87, 88] by a value VGABAA,i I halo,i , where 4mVcm2 /μA . The second 780 process has only a small quantitative effect on the spiking responses of cortical neurons. Note 781 that non-expressing neurons can be interpreted as non-FS neurons, or L4 FS neurons with low 782 halorhodopsin expression levels. 783 784 Numerical methods. Simulations were performed using the fourth-order Runge-Kutta method 785 with time step Δt = 0.05 ms. Simulations with smaller Δt reveal similar statistics of neuronal 786 firing patterns, such as spike rates ν (averaged over many whisking cycles), touch responses R 787 or levels of synchrony. Differences between individual voltage time courses, however, 788 diverged over large integration time interval This divergence is expected because of the chaotic 789 nature of the sparse network dynamics [36, 37, 89]. Statistics were computed after removing a 790 transient of 0.5 s, over 60 s for panels showing dynamical properties of one realization (Fig. 791 4a,c,d, 6d-i, 9a-f ) and over 5.5 s for panel showing dependence on parameters (Fig. 4b, 5, 6b,c, 792 7, 8, 9g,h, 10) where each data point is computed by averaging over 10 network realizations. 793 794 Acknowledgements 795 We thank David Kleinfeld, Rasmus Petersen, Rony Azouz, for helpful discussions, Amy Hu, 796 Brian Barbarits, Sun Wei-Lung, Tim Harris for help with experiments. Gutnisky et al 11/28/2016 797 Conflict of interest 798 The authors declare no conflict of interest. 799 800 801 39 Gutnisky et al 802 References 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 11/28/2016 40 Hubel, D.H. and T.N. Wiesel, Receptive fields, binocular interaction and functional architecture in the cat's visual cortex. J. Physiol. (London), 1962. 160: p. 106-154. Simons, D.J. and G.E. Carvell, Thalamocortical response transformation in the rat vibrissa/barrel system. J Neurophysiol, 1989. 61(2): p. 311-30. Brecht, M. and B. Sakmann, Dynamic representation of whisker deflection by synaptic potentials in spiny stellate and pyramidal cells in the barrels and septa of layer 4 rat somatosensory cortex. J Physiol, 2002. 543(Pt 1): p. 49-70. Brecht, M. and B. Sakmann, Whisker maps of neuronal subclasses of the rat ventral posterior medial thalamus, identified by whole-cell voltage recording and morphological reconstruction. J Physiol, 2002. 538(Pt 2): p. 495-515. Stanley, G.B., et al., Visual orientation and directional selectivity through thalamic synchrony. J Neurosci, 2012. 32(26): p. 9073-88. Reid, R.C. and J.M. Alonso, Specificity of monosynaptic connections from thalamus to visual cortex. Nature, 1995. 378(6554): p. 281-4. Sun, W., et al., Thalamus provides layer 4 of primary visual cortex with orientationand direction-tuned inputs. Nat Neurosci, 2016. 19(2): p. 308-15. Kwegyir-Afful, E.E., H.T. Kyriazi, and D.J. Simons, Weaker feedforward inhibition accounts for less pronounced thalamocortical response transformation in mouse vs. rat barrels. J Neurophysiol, 2013. 110(10): p. 2378-92. Yu, C., et al., Coding of object location in the vibrissal thalamocortical system. Cereb Cortex, 2015. 25(3): p. 563-77. Yarbus, A.L., Eye Movements and Vision. 1967, New York: Plenum Press. Gibson, J.J., Observations on active touch. Psychol Rev, 1962. 69: p. 477-91. Vincent, S.B., The function of vibrissae in the behavior of the white rat. Behavior Monographs, 1912. 1(5): p. 1-82. Diamond, M.E., et al., 'Where' and 'what' in the whisker sensorimotor system. Nat Rev Neurosci, 2008. 9(8): p. 601-12. Sofroniew, N.J. and K. Svoboda, Whisking. Curr Biol, 2015. 25(4): p. R137-40. Hulliger, M., et al., The responses of afferent fibres from the glabrous skin of the hand during voluntary finger movements in man. J Physiol, 1979. 291: p. 233-49. Fee, M.S., P.P. Mitra, and D. Kleinfeld, Central versus peripheral determinants of patterned spike activity in rat vibrissa cortex during whisking. J Neurophysiol, 1997. 78(2): p. 1144-9. Khatri, V., et al., Whisking in air: encoding of kinematics by VPM neurons in awake rats. Somatosens Mot Res, 2010. 27(3): p. 111-20. Moore, J.D., et al., Vibrissa Self-Motion and Touch Are Reliably Encoded along the Same Somatosensory Pathway from Brainstem through Thalamus. PLoS Biol, 2015. 13(9): p. e1002253. Yu, C., et al., Parallel thalamic pathways for whisking and touch signals in the rat. PLoS biology, 2006. 4(5): p. e124. Poulet, J.F., et al., Thalamic control of cortical states. Nature neuroscience, 2012. 15(3): p. 370-2. Gutnisky et al 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 11/28/2016 41 Urbain, N., et al., Whisking-Related Changes in Neuronal Firing and Membrane Potential Dynamics in the Somatosensory Thalamus of Awake Mice. Cell Rep, 2015. 13(4): p. 647-56. Papakostopoulos, D., R. Cooper, and H.J. Crow, Inhibition of cortical evoked potentials and sensation by self-initiated movement in man. Nature, 1975. 258(5533): p. 321-4. Giblin, D.R., Somatosensory Evoked Potentials in Healthy Subjects and in Patients with Lesions of the Nervous System. Ann N Y Acad Sci, 1964. 112: p. 93-142. Chapman, C.E., Active versus passive touch: factors influencing the transmission of somatosensory signals to primary somatosensory cortex. Can J Physiol Pharmacol, 1994. 72(5): p. 558-70. Ma, Y., H. Hu, and A. Agmon, Short-term plasticity of unitary inhibitory-to-inhibitory synapses depends on the presynaptic interneuron subtype. The Journal of neuroscience : the official journal of the Society for Neuroscience, 2012. 32(3): p. 9838. Lefort, S., et al., The excitatory neuronal network of the C2 barrel column in mouse primary somatosensory cortex. Neuron, 2009. 61(2): p. 301-16. Hooks, B.M., et al., Laminar analysis of excitatory local circuits in vibrissal motor and sensory cortical areas. PLoS Biol, 2011. 9(1): p. e1000572. Bureau, I., F. von Saint Paul, and K. Svoboda, Interdigitated Paralemniscal and Lemniscal Pathways in the Mouse Barrel Cortex. PLoS Biol, 2006. 4(12): p. e382. Kinnischtzke, A.K., D.J. Simons, and E.E. Fanselow, Motor Cortex Broadly Engages Excitatory and Inhibitory Neurons in Somatosensory Barrel Cortex. Cerebral Cortex, 2014. 24(8): p. 2237-2248. Porter, J.T., C.K. Johnson, and A. Agmon, Diverse types of interneurons generate thalamus-evoked feedforward inhibition in the mouse barrel cortex. J Neurosci, 2001. 21(8): p. 2699-710. Bruno, R.M. and D.J. Simons, Feedforward mechanisms of excitatory and inhibitory cortical receptive fields. The Journal of neuroscience, 2002. 22(24): p. 10966-75. Cruikshank, S.J., T.J. Lewis, and B.W. Connors, Synaptic basis for intense thalamocortical activation of feedforward inhibitory cells in neocortex. Nat Neurosci, 2007. 10(4): p. 462-8. Gabernet, L., et al., Somatosensory integration controlled by dynamic thalamocortical feed-forward inhibition. Neuron, 2005. 48(2): p. 315-27. Hires, S.A., et al., Low-noise encoding of active touch by layer 4 in the somatosensory cortex. Elife, 2015. 4. Yu, J., et al., Layer 4 fast-spiking interneurons filter thalamocortical signals during active somatosensation. Nat Neurosci, 2016. van Vreeswijk, C. and H. Sompolinsky, Chaos in neuronal networks with balanced excitatory and inhibitory activity. Science, 1996. 274(5293): p. 1724-6. van Vreeswijk, C. and H. Sompolinsky, Chaotic balanced state in a model of cortical circuits. Neural computation, 1998. 10(6): p. 1321-71. Pehlevan, C. and H. Sompolinsky, Selectivity and sparseness in randomly connected balanced networks. PLoS One, 2014. 9(2): p. e89992. Tsodyks, M.V., et al., Paradoxical effects of external modulation of inhibitory interneurons. J Neurosci, 1997. 17(11): p. 4382-8. Gutnisky et al 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 11/28/2016 42 Ozeki, H., et al., Inhibitory stabilization of the cortical network underlies visual surround suppression. Neuron, 2009. 62(4): p. 578-92. Knutsen, P.M., M. Pietr, and E. Ahissar, Haptic object localization in the vibrissal system: behavior and performance. J Neurosci, 2006. 26(33): p. 8451-64. O'Connor, D.H., et al., Vibrissa-based object localization in head-fixed mice. J Neurosci, 2010. 30(5): p. 1947-67. Clack, N.G., et al., Automated tracking of whiskers in videos of head fixed rodents. PLoS computational biology, 2012. 8(7): p. e1002591. Pammer, L., et al., The Mechanical Variables Underlying Object Localization along the Axis of the Whisker. J Neurosci, 2013. 33(16): p. 6726-41. O'Connor, D.H., et al., Neural coding during active somatosensation revealed using illusory touch. Nat Neurosci, 2013. 16(7): p. 958-65. Bruno, R.M. and B. Sakmann, Cortex is driven by weak but synchronously active thalamocortical synapses. Science, 2006. 312(5780): p. 1622-7. Han, X. and E.S. Boyden, Multiple-color optical activation, silencing, and desynchronization of neural activity, with single-spike temporal resolution. PLoS ONE, 2007. 2: p. e299. Zhang, F., et al., Multimodal fast optical interrogation of neural circuitry. Nature, 2007. 446(7136): p. 633-9. Beierlein, M., J.R. Gibson, and B.W. Connors, Two dynamically distinct inhibitory networks in layer 4 of the neocortex. J Neurophysiol, 2003. 90(5): p. 2987-3000. Feldmeyer, D., A. Roth, and B. Sakmann, Monosynaptic connections between pairs of spiny stellate cells in layer 4 and pyramidal cells in layer 5A indicate that lemniscal and paralemniscal afferent pathways converge in the infragranular somatosensory cortex. The Journal of neuroscience : the official journal of the Society for Neuroscience, 2005. 25(13): p. 3423-31. Hansel, D. and C. van Vreeswijk, The mechanism of orientation selectivity in primary visual cortex without a functional map. J Neurosci, 2012. 32(12): p. 4049-64. Wang, X.J. and G. Buzsaki, Gamma oscillation by synaptic inhibition in a hippocampal interneuronal network model. J Neurosci, 1996. 16(20): p. 6402-13. Hull, C., J.S. Isaacson, and M. Scanziani, Postsynaptic mechanisms govern the differential excitation of cortical neurons by thalamic inputs. The Journal of neuroscience : the official journal of the Society for Neuroscience, 2009. 29(28): p. 9127-36. Pinto, D.J., et al., Cortical damping: analysis of thalamocortical response transformations in rodent barrel cortex. Cereb Cortex, 2003. 13(1): p. 33-44. Wilent, W.B. and D. Contreras, Dynamics of excitation and inhibition underlying stimulus selectivity in rat somatosensory cortex. Nat Neurosci, 2005. 8(10): p. 1364-70. Kremkow, J., et al., Functional consequences of correlated excitatory and inhibitory conductances in cortical networks. J Comput Neurosci, 2010. 28(3): p. 579-94. Pinto, D.J., J.C. Brumberg, and D.J. Simons, Circuit dynamics and coding strategies in rodent somatosensory cortex. J Neurophysiol, 2000. 83(3): p. 1158-66. Huber, D., et al., Sparse optical microstimulation in barrel cortex drives learned behaviour in freely moving mice. Nature, 2008. 451(7174): p. 61-4. Gutnisky et al 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 11/28/2016 43 Rubin, D.B., S.D. Van Hooser, and K.D. Miller, The stabilized supralinear network: a unifying circuit motif underlying multi-input integration in sensory cortex. Neuron, 2015. 85(2): p. 402-17. O'Connor, D.H., et al., Neural activity in barrel cortex underlying vibrissa-based object localization in mice. Neuron, 2010. 67(6): p. 1048-61. Agmon, A. and B.W. Connors, Thalamocortical responses of mouse somatosensory (barrel) cortex in vitro. Neuroscience, 1991. 41(2-3): p. 365-79. Wehr, M. and A.M. Zador, Balanced inhibition underlies tuning and sharpens spike timing in auditory cortex. Nature, 2003. 426(6965): p. 442-6. DeWeese, M.R., M. Wehr, and A.M. Zador, Binary spiking in auditory cortex. J Neurosci, 2003. 23(21): p. 7940-9. Hayut, I., et al., LTS and FS inhibitory interneurons, short-term synaptic plasticity, and cortical circuit dynamics. PLoS Comput Biol, 2011. 7(10): p. e1002248. Markram, H., et al., Reconstruction and Simulation of Neocortical Microcircuitry. Cell, 2015. 163(2): p. 456-92. Traub, R.D., R. Miles, and R.K. Wong, Model of the origin of rhythmic population oscillations in the hippocampal slice. Science, 1989. 243(4896): p. 1319-25. Eliasmith, C. and O. Trujillo, The use and abuse of large-scale brain models. Curr Opin Neurobiol, 2014. 25: p. 1-6. Marder, E. and J.M. Goaillard, Variability, compensation and homeostasis in neuron and network function. Nat Rev Neurosci, 2006. 7(7): p. 563-74. Shriki, O., D. Hansel, and H. Sompolinsky, Rate models for conductance-based cortical neuronal networks. Neural Comput, 2003. 15(8): p. 1809-41. Reinhold, K., A.D. Lien, and M. Scanziani, Distinct recurrent versus afferent dynamics in cortical visual processing. Nat Neurosci, 2015. 18(12): p. 1789-97. Bruno, R.M., Synchrony in sensation. Current opinion in neurobiology, 2011. 21(5): p. 701-8. Reyes, A.D., Synchrony-dependent propagation of firing rate in iteratively constructed networks in vitro. Nat Neurosci, 2003. 6(6): p. 593-9. Lerchner, A., et al., Mean field theory for a balanced hypercolumn model of orientation selectivity in primary visual cortex. Network, 2006. 17(2): p. 131-50. Hansel, D. and G. Mato, Short-term plasticity explains irregular persistent activity in working memory tasks. J Neurosci, 2013. 33(1): p. 133-49. Ahmadian, Y., D.B. Rubin, and K.D. Miller, Analysis of the stabilized supralinear network. Neural Comput, 2013. 25(8): p. 1994-2037. Persi, E., et al., Power-law input-output transfer functions explain the contrastresponse and tuning properties of neurons in visual cortex. PLoS Comput Biol, 2011. 7(2): p. e1001078. Birdwell, J.A., et al., Biomechanical models for radial distance determination by the rat vibrissal system. J Neurophysiol, 2007. 98(4): p. 2439-55. Song, S., et al., Highly nonrandom features of synaptic connectivity in local cortical circuits. PLoS Biol, 2005. 3(3): p. 1-13. Ma, Y., et al., Distinct subtypes of somatostatin-containing neocortical interneurons revealed in transgenic mice. J Neurosci, 2006. 26(19): p. 5069-82. Oh, S.W., et al., A mesoscale connectome of the mouse brain. Nature, 2014. 508(7495): p. 207-14. Gutnisky et al 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 11/28/2016 44 Petreanu, L., et al., The subcellular organization of neocortical excitatory connections. Nature, 2009. 457(7233): p. 1142-1145. Jiang, X., et al., Principles of connectivity among morphologically defined cell types in adult neocortex. Science, 2015. 350(6264): p. aac9462. Ferster, D., S. Chung, and H. Wheat, Orientation selectivity of thalamic input to simple cells of cat visual cortex. Nature, 1996. 380: p. 249-252. Golomb, D., et al., On temporal codes and the spatiotemporal response of neurons in the lateral geniculate nucleus. J Neurophysiol, 1994. 72(6): p. 2990-3003. Golomb, D. and D. Hansel, The number of synaptic inputs and the synchrony of large, sparse neuronal networks. Neural Comput, 2000. 12(5): p. 1095-139. Varshney, L.R., P.J. Sjostrom, and D.B. Chklovskii, Optimal information storage in noisy synapses under resource constraints. Neuron, 2006. 52(3): p. 409-23. Alfonsa, H., et al., The contribution of raised intraneuronal chloride to epileptic network activity. J Neurosci, 2015. 35(20): p. 7715-26. Raimondo, J.V., et al., Optogenetic silencing strategies differ in their effects on inhibitory synaptic transmission. Nat Neurosci, 2012. 15(8): p. 1102-4. Zillmer, R., N. Brunel, and D. Hansel, Very long transients, irregular firing, and chaotic dynamics in networks of randomly connected inhibitory integrate-and-fire neurons. Phys Rev E Stat Nonlin Soft Matter Phys, 2009. 79(3 Pt 1): p. 031909. Fanselow, E.E., et al., Thalamic bursting in rats during different awake behavioral states. Proc Natl Acad Sci U S A, 2001. 98(26): p. 15330-5. 1002 1003 1004 Supporting Information 1005 1006 S1 Fig. Example neurons recorded in VPM during whisker-based object localization. (a) 1007 Chronic silicon probe recordings in VPM. (b,c) Two example trials from the same neuron 1008 showing increases in activity with whisker movements without touch. Blue line, whisker 1009 azimuthal angle. Black ticks, spikes. (d,e) Two example trials showing increases in spike rate 1010 after touch. (f) Touch rasters (black dots; sorted by last touch in a trial). The magenta line 1011 shows when the pole was moved within reach of the whiskers. The green line represents the 1012 last touch in each trial. (g) Whisker movement amplitude. Red rectangle, epoch of high 1013 whisker movement amplitude and high spike rate (same as in h). Orange rectangle, epoch of Gutnisky et al 11/28/2016 45 1014 low amplitude and low spike rate (same as in h). (h) Spike raster for an example neuron in 1015 VPM (blue dots, spikes). The black arrow in g-h indicates onset of whisker twitching [90]. The 1016 twitching is triggered by an auditory signal generated by the brief activation of a shutter. (i) 1017 Peri-stimulus-time-histogram (PSTH) showing response for touch. (j) Spike rate as a function 1018 of whisker movement amplitude. (k-o) Same as f-j for another example VPM neuron with low 1019 baseline spike rate. (p-t) Same as f-j for another example VPM neuron that does not show 1020 modulation with whisking amplitude. 1021 1022 S2 Fig. Example neurons across the thalamocortical circuit. (a) Touch raster (black dots; 1023 sorted by last touch in a trial). The magenta line shows when the pole was moved within reach 1024 of the whiskers. The green line represents the last touch in each trial. (b) Whisker movement 1025 amplitude. Red rectangle, epoch of high whisker movement amplitude and high spike rate. 1026 Orange rectangle, epoch of low amplitude and low spike rate. (c) Spike raster of an example 1027 neuron in VPM (blue dots, spikes). Red and orange regions of interest correspond to B. (d) 1028 PSTH showing response for touch. (e) Spike rate as a function of whisker movement amplitude. 1029 (f-j) Same as a-e for a L4 FS neuron. (k-o) Same as a-f for a L4 excitatory neuron. The few 1030 spikes that occur during whisking are phase-locked to movement. 1031