Mapping Functional Connectivity from the Dorsal Cortex to the Thalamus

•A high throughput pipeline to map functional corticothalamic connectivity•Inactivation of the cortex rapidly reduces thalamic activity•Functional corticothalamic connectivity shows topographical organization•Some thalamic neurons receive convergent inputs from widespread cortical regions Neural activity in the corticothalamic network is crucial for sensation, memory, decision, and action. Nevertheless, a systematic characterization of corticothalamic functional connectivity has not been achieved. Here, we developed a high throughput method to systematically map functional connections from the dorsal cortex to the thalamus in awake mice by combing optogenetic inactivation with multi-channel recording. Cortical inactivation resulted in a rapid reduction of thalamic activity, revealing topographically organized corticothalamic excitatory inputs. Cluster analysis showed that groups of neurons within individual thalamic nuclei exhibited distinct dynamics. The effects of inactivation evolved with time and were modulated by behavioral states. Furthermore, we found that a subset of thalamic neurons received convergent inputs from widespread cortical regions. Our results present a framework for collecting, analyzing, and presenting large electrophysiological datasets with region-specific optogenetic perturbations and serve as a foundation for further investigation of information processing in the corticothalamic pathway. Neural activity in the corticothalamic network is crucial for sensation, memory, decision, and action. Nevertheless, a systematic characterization of corticothalamic functional connectivity has not been achieved. Here, we developed a high throughput method to systematically map functional connections from the dorsal cortex to the thalamus in awake mice by combing optogenetic inactivation with multi-channel recording. Cortical inactivation resulted in a rapid reduction of thalamic activity, revealing topographically organized corticothalamic excitatory inputs. Cluster analysis showed that groups of neurons within individual thalamic nuclei exhibited distinct dynamics. The effects of inactivation evolved with time and were modulated by behavioral states. Furthermore, we found that a subset of thalamic neurons received convergent inputs from widespread cortical regions. Our results present a framework for collecting, analyzing, and presenting large electrophysiological datasets with region-specific optogenetic perturbations and serve as a foundation for further investigation of information processing in the corticothalamic pathway. The cerebral cortex consists of dozens of functionally distinct regions that form complex networks through their dense corticocortical connections (Felleman and Van Essen, 1991Felleman D.J. Van Essen D.C. Distributed hierarchical processing in the primate cerebral cortex.Cereb. Cortex. 1991; 1: 1-47Crossref PubMed Scopus (5141) Google Scholar; Glasser et al., 2016Glasser M.F. Coalson T.S. Robinson E.C. Hacker C.D. Harwell J. Yacoub E. Ugurbil K. Andersson J. Beckmann C.F. Jenkinson M. et al.A multi-modal parcellation of human cerebral cortex.Nature. 2016; 536: 171-178Crossref PubMed Scopus (1394) Google Scholar; Harris et al., 2019Harris J.A. Mihalas S. Hirokawa K.E. Whitesell J.D. Choi H. Bernard A. Bohn P. Caldejon S. Casal L. Cho A. et al.Hierarchical organization of cortical and thalamic connectivity.Nature. 2019; 575: 195-202Crossref PubMed Scopus (83) Google Scholar; Oh et al., 2014Oh S.W. Harris J.A. Ng L. Winslow B. Cain N. Mihalas S. Wang Q. Lau C. Kuan L. Henry A.M. et al.A mesoscale connectome of the mouse brain.Nature. 2014; 508: 207-214Crossref PubMed Scopus (1072) Google Scholar; Zingg et al., 2014Zingg B. Hintiryan H. Gou L. Song M.Y. Bay M. Bienkowski M.S. Foster N.N. Yamashita S. Bowman I. Toga A.W. Dong H.W. Neural networks of the mouse neocortex.Cell. 2014; 156: 1096-1111Abstract Full Text Full Text PDF PubMed Scopus (353) Google Scholar). Additionally, the cortex is heavily connected to the thalamus through reciprocal connections and cortico-cortico-thalamic pathways (Harris et al., 2019Harris J.A. Mihalas S. Hirokawa K.E. Whitesell J.D. Choi H. Bernard A. Bohn P. Caldejon S. Casal L. Cho A. et al.Hierarchical organization of cortical and thalamic connectivity.Nature. 2019; 575: 195-202Crossref PubMed Scopus (83) Google Scholar; Hunnicutt et al., 2014Hunnicutt B.J. Long B.R. Kusefoglu D. Gertz K.J. Zhong H. Mao T. A comprehensive thalamocortical projection map at the mesoscopic level.Nat. Neurosci. 2014; 17: 1276-1285Crossref PubMed Scopus (68) Google Scholar; Oh et al., 2014Oh S.W. Harris J.A. Ng L. Winslow B. Cain N. Mihalas S. Wang Q. Lau C. Kuan L. Henry A.M. et al.A mesoscale connectome of the mouse brain.Nature. 2014; 508: 207-214Crossref PubMed Scopus (1072) Google Scholar; Scannell et al., 1999Scannell J.W. Burns G.A. Hilgetag C.C. O’Neil M.A. Young M.P. The connectional organization of the cortico-thalamic system of the cat.Cereb. Cortex. 1999; 9: 277-299Crossref PubMed Scopus (288) Google Scholar), which are crucial for sensory processing, perception, working memory, decision making, motor planning, and execution of actions (Alexander and Fuster, 1973Alexander G.E. Fuster J.M. Effects of cooling prefrontal cortex on cell firing in the nucleus medialis dorsalis.Brain Res. 1973; 61: 93-105Crossref PubMed Scopus (70) Google Scholar; Bolkan et al., 2017Bolkan S.S. Stujenske J.M. Parnaudeau S. Spellman T.J. Rauffenbart C. Abbas A.I. Harris A.Z. Gordon J.A. Kellendonk C. Thalamic projections sustain prefrontal activity during working memory maintenance.Nat. Neurosci. 2017; 20: 987-996Crossref PubMed Scopus (183) Google Scholar; Guo et al., 2017Guo Z.V. Inagaki H.K. Daie K. Druckmann S. Gerfen C.R. Svoboda K. Maintenance of persistent activity in a frontal thalamocortical loop.Nature. 2017; 545: 181-186Crossref PubMed Scopus (174) Google Scholar; Schmitt et al., 2017Schmitt L.I. Wimmer R.D. Nakajima M. Happ M. Mofakham S. Halassa M.M. Thalamic amplification of cortical connectivity sustains attentional control.Nature. 2017; 545: 219-223Crossref PubMed Scopus (228) Google Scholar; Sherman, 2016Sherman S.M. Thalamus plays a central role in ongoing cortical functioning.Nat. Neurosci. 2016; 19: 533-541Crossref PubMed Scopus (301) Google Scholar; Vanduffel et al., 1997Vanduffel W. Payne B.R. Lomber S.G. Orban G.A. Functional impact of cerebral connections.Proc. Natl. Acad. Sci. USA. 1997; 94: 7617-7620Crossref PubMed Scopus (71) Google Scholar). How the cortex interacts with the thalamus to implement its functions remains poorly understood. A crucial step is to understand how individual cortical areas regulate thalamic activity through their corticothalamic projections. Corticothalamic projections have been demonstrated to strongly influence thalamic activity in various mammalian species, including primates, cats, and rodents (Alexander and Fuster, 1973Alexander G.E. Fuster J.M. Effects of cooling prefrontal cortex on cell firing in the nucleus medialis dorsalis.Brain Res. 1973; 61: 93-105Crossref PubMed Scopus (70) Google Scholar; Bolkan et al., 2017Bolkan S.S. Stujenske J.M. Parnaudeau S. Spellman T.J. Rauffenbart C. Abbas A.I. Harris A.Z. Gordon J.A. Kellendonk C. Thalamic projections sustain prefrontal activity during working memory maintenance.Nat. Neurosci. 2017; 20: 987-996Crossref PubMed Scopus (183) Google Scholar; Chang, 1950Chang H.-T. The repetitive discharges of corticothalamic reverberating circuit.J. Neurophysiol. 1950; 13: 235-257Crossref PubMed Scopus (73) Google Scholar; Guo et al., 2017Guo Z.V. Inagaki H.K. Daie K. Druckmann S. Gerfen C.R. Svoboda K. Maintenance of persistent activity in a frontal thalamocortical loop.Nature. 2017; 545: 181-186Crossref PubMed Scopus (174) Google Scholar; Head and Holmes, 1912Head H. Holmes G. Sensory disturbances from cerebral lesions.Brain. 1912; 34: 102-254Crossref Google Scholar; Schmitt et al., 2017Schmitt L.I. Wimmer R.D. Nakajima M. Happ M. Mofakham S. Halassa M.M. Thalamic amplification of cortical connectivity sustains attentional control.Nature. 2017; 545: 219-223Crossref PubMed Scopus (228) Google Scholar; Theyel et al., 2010Theyel B.B. Llano D.A. Sherman S.M. The corticothalamocortical circuit drives higher-order cortex in the mouse.Nat. Neurosci. 2010; 13: 84-88Crossref PubMed Scopus (212) Google Scholar; Vanduffel et al., 1997Vanduffel W. Payne B.R. Lomber S.G. Orban G.A. Functional impact of cerebral connections.Proc. Natl. Acad. Sci. USA. 1997; 94: 7617-7620Crossref PubMed Scopus (71) Google Scholar). Nevertheless, there is no systematic study to map the causal contribution of cortical activity to thalamic activity (defined here as functional corticothalamic connectivity [FCC]). It remains challenging to create a comprehensive functional corticothalamic map from individual cortical regions to each thalamic nucleus because the cortex and thalamus have extensive projections to each other (Harris et al., 2019Harris J.A. Mihalas S. Hirokawa K.E. Whitesell J.D. Choi H. Bernard A. Bohn P. Caldejon S. Casal L. Cho A. et al.Hierarchical organization of cortical and thalamic connectivity.Nature. 2019; 575: 195-202Crossref PubMed Scopus (83) Google Scholar; Hunnicutt et al., 2014Hunnicutt B.J. Long B.R. Kusefoglu D. Gertz K.J. Zhong H. Mao T. A comprehensive thalamocortical projection map at the mesoscopic level.Nat. Neurosci. 2014; 17: 1276-1285Crossref PubMed Scopus (68) Google Scholar; Oh et al., 2014Oh S.W. Harris J.A. Ng L. Winslow B. Cain N. Mihalas S. Wang Q. Lau C. Kuan L. Henry A.M. et al.A mesoscale connectome of the mouse brain.Nature. 2014; 508: 207-214Crossref PubMed Scopus (1072) Google Scholar; Scannell et al., 1999Scannell J.W. Burns G.A. Hilgetag C.C. O’Neil M.A. Young M.P. The connectional organization of the cortico-thalamic system of the cat.Cereb. Cortex. 1999; 9: 277-299Crossref PubMed Scopus (288) Google Scholar), and thus, inhibiting one cortical region will inevitably affect a wide range of cortical and thalamic areas at a longer timescale. Physical lesions, pharmacological manipulations, and cooling methods do not have sufficient temporal resolution to differentiate the direct and indirect effects (Hikosaka and Wurtz, 1985Hikosaka O. Wurtz R.H. Modification of saccadic eye movements by GABA-related substances. I. Effect of muscimol and bicuculline in monkey superior colliculus.J. Neurophysiol. 1985; 53: 266-291Crossref PubMed Scopus (450) Google Scholar; Long and Fee, 2008Long M.A. Fee M.S. Using temperature to analyse temporal dynamics in the songbird motor pathway.Nature. 2008; 456: 189-194Crossref PubMed Scopus (289) Google Scholar; Ponce et al., 2008Ponce C.R. Lomber S.G. Born R.T. Integrating motion and depth via parallel pathways.Nat. Neurosci. 2008; 11: 216-223Crossref PubMed Scopus (86) Google Scholar). Thus, dissecting the direct effects will require a method to perturb and monitor neural activity with milliseconds of temporal resolution. Next, monitoring thalamic activity requires recording neural activity from deep brain areas, necessitating a method to accurately reconstruct the electrode locations in order to assign the recorded neurons to individual thalamic nuclei that are densely packed. Furthermore, both the cortex and thalamus have dozens of distinct areas spanning millimeters in space, requiring a high throughput method to perturb neural activity in many cortical regions repeatedly while simultaneously recording thalamic activity from multiple thalamic sites. Lastly, collecting, analyzing, and presenting large electrophysiological datasets from many thalamic nuclei remains challenging. Here, we developed a high throughput pipeline by combining optogenetic local cortical inactivation with thalamic multi-channel recording to systematically map FCC in awake mice. Optogenetic inactivation allowed us to reversibly silence ∼1 mm3 cortical tissue with milliseconds of temporal resolution (Deisseroth, 2015Deisseroth K. Optogenetics: 10 years of microbial opsins in neuroscience.Nat. Neurosci. 2015; 18: 1213-1225Crossref PubMed Scopus (582) Google Scholar; Guo et al., 2014bGuo Z.V. Li N. Huber D. Ophir E. Gutnisky D. Ting J.T. Feng G. Svoboda K. Flow of cortical activity underlying a tactile decision in mice.Neuron. 2014; 81: 179-194Abstract Full Text Full Text PDF PubMed Scopus (294) Google Scholar). A galvo scanning system enabled targeting 34 cortical regions (∼50% of neocortex) repeatedly in an addressable coordinate system (Guo et al., 2014bGuo Z.V. Li N. Huber D. Ophir E. Gutnisky D. Ting J.T. Feng G. Svoboda K. Flow of cortical activity underlying a tactile decision in mice.Neuron. 2014; 81: 179-194Abstract Full Text Full Text PDF PubMed Scopus (294) Google Scholar). Silicon probes with multi-recording channels collected neural activity from dozens of neurons at one thalamic location. We imaged the recorded brains with electrode tracks at micrometer resolution and aligned the brains to the Common Coordinate Framework (CCF) (Wang et al., 2020Wang Q. Ding S.L. Li Y. Royall J. Feng D. Lesnar P. Graddis N. Naeemi M. Facer B. Ho A. et al.The Allen Mouse Brain Common Coordinate Framework: A 3D Reference Atlas.Cell. 2020; 181: 936-953.e20Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar). By combining information from CCF, we analyzed the dynamic response of thalamic neurons to inactivation of one of 34 cortical regions. Our results present a framework for collecting, analyzing, and presenting region-specific optogenetic perturbations combined with large electrophysiological datasets. Inactivation of the sensory, motor, and association cortices suppressed thalamic activity with a short latency. FCC from these cortices showed topographical organization, with the anterior cortex preferentially projecting to the ventral and medial part of individual thalamic nuclei and the lateral part of the cortex preferentially projecting to the lateral and posterior part of individual thalamic nuclei. Cluster analysis revealed that different sets of thalamic neurons received distinct patterns of cortical inputs and exhibited different temporal dynamics following cortical inactivation. Interestingly, a subset of thalamic neurons received convergent innervations from the sensory, motor, and association cortices, implying that these neurons might function to integrate diverse cortical information. Given the widespread corticothalamic, thalamocortical, and corticocortical connections (Felleman and Van Essen, 1991Felleman D.J. Van Essen D.C. Distributed hierarchical processing in the primate cerebral cortex.Cereb. Cortex. 1991; 1: 1-47Crossref PubMed Scopus (5141) Google Scholar; Harris et al., 2019Harris J.A. Mihalas S. Hirokawa K.E. Whitesell J.D. Choi H. Bernard A. Bohn P. Caldejon S. Casal L. Cho A. et al.Hierarchical organization of cortical and thalamic connectivity.Nature. 2019; 575: 195-202Crossref PubMed Scopus (83) Google Scholar; Hunnicutt et al., 2014Hunnicutt B.J. Long B.R. Kusefoglu D. Gertz K.J. Zhong H. Mao T. A comprehensive thalamocortical projection map at the mesoscopic level.Nat. Neurosci. 2014; 17: 1276-1285Crossref PubMed Scopus (68) Google Scholar; Oh et al., 2014Oh S.W. Harris J.A. Ng L. Winslow B. Cain N. Mihalas S. Wang Q. Lau C. Kuan L. Henry A.M. et al.A mesoscale connectome of the mouse brain.Nature. 2014; 508: 207-214Crossref PubMed Scopus (1072) Google Scholar; Scannell et al., 1999Scannell J.W. Burns G.A. Hilgetag C.C. O’Neil M.A. Young M.P. The connectional organization of the cortico-thalamic system of the cat.Cereb. Cortex. 1999; 9: 277-299Crossref PubMed Scopus (288) Google Scholar; Zingg et al., 2014Zingg B. Hintiryan H. Gou L. Song M.Y. Bay M. Bienkowski M.S. Foster N.N. Yamashita S. Bowman I. Toga A.W. Dong H.W. Neural networks of the mouse neocortex.Cell. 2014; 156: 1096-1111Abstract Full Text Full Text PDF PubMed Scopus (353) Google Scholar), our results serve as a foundation for further investigation of corticothalamic and thalamocortical functions. We inactivated individual cortical regions while simultaneously recording from multiple thalamic locations to probe FCC (Figure 1A). Optogenetic inactivation was achieved by photostimulating a red-shifted cruxhalorhodopsin (Jaws) in pyramidal neurons in a triple transgenic mouse line Emx1-IRES-Cre; Camk2a-tTA; Ai79 (Figure 1B) (Chuong et al., 2014Chuong A.S. Miri M.L. Busskamp V. Matthews G.A. Acker L.C. Sørensen A.T. Young A. Klapoetke N.C. Henninger M.A. Kodandaramaiah S.B. et al.Noninvasive optical inhibition with a red-shifted microbial rhodopsin.Nat. Neurosci. 2014; 17: 1123-1129Crossref PubMed Scopus (295) Google Scholar; Madisen et al., 2015Madisen L. Garner A.R. Shimaoka D. Chuong A.S. Klapoetke N.C. Li L. van der Bourg A. Niino Y. Egolf L. Monetti C. et al.Transgenic mice for intersectional targeting of neural sensors and effectors with high specificity and performance.Neuron. 2015; 85: 942-958Abstract Full Text Full Text PDF PubMed Scopus (477) Google Scholar). To efficiently map FCC, we employed a galvo scanning system to target photostimuli repeatedly in a random-access manner to 34 cortical locations (covering ∼50% of the neocortex; Figures 1A, 1C, and S1F) (Guo et al., 2014bGuo Z.V. Li N. Huber D. Ophir E. Gutnisky D. Ting J.T. Feng G. Svoboda K. Flow of cortical activity underlying a tactile decision in mice.Neuron. 2014; 81: 179-194Abstract Full Text Full Text PDF PubMed Scopus (294) Google Scholar). In parallel to inactivation, we recorded extracellular activity with a 64-channel silicon probe across multiple thalamic nuclei. We advanced the silicon probe to record from a new set of thalamic locations in the same mouse between sessions (∼5 sessions per probe insertion; Figure 2A). Combining the two approaches effectively increased the throughput ∼200 times, compared with the approach using a single cortical region inactivation and a single thalamic location recording, to enable efficiently survey of FCC (Figures 1C and 1D).Figure 23D Reconstruction of Electrode LocationsShow full caption(A) The pipeline for 3D reconstruction of electrode tracks. Blue crosses indicate locations of electrode tips (4-shank probe) from 5 sessions. Scale bar, 1 mm. Arrow in the inset indicates the electric lesion site. Scale bar, 200 μm.(B) Error of the boundary between the reconstructed thalamus and the model thalamus from the CCF. Top: schematic of the twelve directions used for comparison. Gray shadow indicates boundary variability of traced thalamus mask. Bottom: polar plot of the averaged error (the inner black line). Gray indicates SEM over animals (n = 43).(C) Schematic of neurophysiology landmarks. Arrow indicates electrode advancing direction.(D) Fraction of narrow spiking neurons near the boundary between VPL-VAL and RT. Thick lines: mean fraction. Light lines: individual mice.(E) Error between the boundary of the 3D reconstructed thalamus and the boundary based on neurophysiology landmarks. Error bar, SEM over animals (n = 6, VAL; n = 5, VPL).(F) Number of recording sites (left, blue) and sessions (right, red) in the thalamus. Slice spacing, 75 μm.See also Figure S2.View Large Image Figure ViewerDownload Hi-res image Download (PPT) (A) The pipeline for 3D reconstruction of electrode tracks. Blue crosses indicate locations of electrode tips (4-shank probe) from 5 sessions. Scale bar, 1 mm. Arrow in the inset indicates the electric lesion site. Scale bar, 200 μm. (B) Error of the boundary between the reconstructed thalamus and the model thalamus from the CCF. Top: schematic of the twelve directions used for comparison. Gray shadow indicates boundary variability of traced thalamus mask. Bottom: polar plot of the averaged error (the inner black line). Gray indicates SEM over animals (n = 43). (C) Schematic of neurophysiology landmarks. Arrow indicates electrode advancing direction. (D) Fraction of narrow spiking neurons near the boundary between VPL-VAL and RT. Thick lines: mean fraction. Light lines: individual mice. (E) Error between the boundary of the 3D reconstructed thalamus and the boundary based on neurophysiology landmarks. Error bar, SEM over animals (n = 6, VAL; n = 5, VPL). (F) Number of recording sites (left, blue) and sessions (right, red) in the thalamus. Slice spacing, 75 μm. See also Figure S2. To characterize the spatial and temporal profile of inactivation, we recorded single units from SSp in awake mice (Figures 1E–1G and S1A–D). The spike widths (from trough to peak) formed a bimodal distribution that allowed the isolation of putative pyramidal neurons (Figure S1A) (Guo et al., 2014bGuo Z.V. Li N. Huber D. Ophir E. Gutnisky D. Ting J.T. Feng G. Svoboda K. Flow of cortical activity underlying a tactile decision in mice.Neuron. 2014; 81: 179-194Abstract Full Text Full Text PDF PubMed Scopus (294) Google Scholar). Inactivation powerfully suppressed neural activity near the laser spot center (activity reduction: 86.3% ± 3.0%, 16 mW laser; 88.9% ± 3.2%, 32 mW, mean ± SEM, laser distance < 0.2 mm; Figures 1E, 1F, S1B, and S1C). The laser power at the dura of the cortex was ∼8 or 16 mW, because the light transmission efficiency was ∼50% (Guo et al., 2014bGuo Z.V. Li N. Huber D. Ophir E. Gutnisky D. Ting J.T. Feng G. Svoboda K. Flow of cortical activity underlying a tactile decision in mice.Neuron. 2014; 81: 179-194Abstract Full Text Full Text PDF PubMed Scopus (294) Google Scholar). This would not produce undesired heating effects, because the laser spot only lasts for 0.4 s (Stujenske et al., 2015Stujenske J.M. Spellman T. Gordon J.A. Modeling the Spatiotemporal Dynamics of Light and Heat Propagation for In Vivo Optogenetics.Cell Rep. 2015; 12: 525-534Abstract Full Text Full Text PDF PubMed Scopus (163) Google Scholar). Inactivation suppressed neural activity with a short latency (1.4 ± 0.3 ms after laser onset, mean ± SEM, laser power 32 mW; Figure S1D). The suppression was potent even in deep cortical layers (>90% in layer 5 [L5] neurons, >50% in L6 neurons, 16mW laser power; Figure 1F). Inactivation was local, having a half-max width of 0.67 or 1.26 mm for L5 and L6 neurons at 16 or 32 mW, respectively (Figure 1G). Based on this spatial profile, inactivation locations were chosen uniformly distributed across the dorsal cortex (mean distance, 0.80 ± 0.08 mm, Kolmogorov-Smirnov test, p > 0.05; Figures S1E–S1G). The mouse thalamus sits deep in the brain, and it is difficult to target specific nuclei due to the uncertainty in determining the recording location. Here, we developed a data collection and analyses pipeline to reconstruct electrode tracks with high precision (Figure 2A; see STAR Methods). First, we painted a thin layer of DiI on the silicon probe before recording, and the DiI fluorescence enabled the visualization of the electrode tracks post hoc. Second, we delivered a brief electric pulse (20 μA for 1 s, 3 to 4 times) to make a small lesion near the tip of the probe. Third, we sectioned and imaged the coronal brain slices from the anterior to posterior thalamus. Fourth, we manually traced the boundary of the thalamus, generated a binary mask, and after tilt correction, aligned it in CCF. During this step, we also marked the electrode tracks and lesion locations. Finally, we stacked the aligned slice masks to generate a 3D thalamus and mapped it to the model thalamus in CCF through rigid transformation (Hunnicutt et al., 2014Hunnicutt B.J. Long B.R. Kusefoglu D. Gertz K.J. Zhong H. Mao T. A comprehensive thalamocortical projection map at the mesoscopic level.Nat. Neurosci. 2014; 17: 1276-1285Crossref PubMed Scopus (68) Google Scholar). We estimated the accuracy of the 3D reconstruction using two complementary approaches. First, we compared the boundary distance between the reconstructed thalamus and model thalamus in CCF. The average distance was 72.4 ± 5.1 μm (mean ± SEM), suggesting that our approach had an error less than 100 μm (Figure 2B). Second, we checked the accuracy by taking advantage of neurophysiological landmarks in the thalamus. Neurons in the reticular nucleus of the thalamus (RT) are GABAergic, and neurons in other thalamic nuclei are mostly glutamatergic. We sorted 382 single units near the boundary of the RT and the neighboring nuclei, including the ventral anterior lateral nucleus (VAL) and the ventral posterolateral nucleus (VPL) (Figure 2C). The spike widths formed a bimodal distribution that allowed to isolate putative RT neurons (and GABAergic axons) from thalamic projection neurons (Figures S2A and S2B) (Barthó et al., 2014Barthó P. Slézia A. Mátyás F. Faradzs-Zade L. Ulbert I. Harris K.D. Acsády L. Ongoing network state controls the length of sleep spindles via inhibitory activity.Neuron. 2014; 82: 1367-1379Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar; Halassa et al., 2014Halassa M.M. Chen Z. Wimmer R.D. Brunetti P.M. Zhao S. Zikopoulos B. Wang F. Brown E.N. Wilson M.A. State-dependent architecture of thalamic reticular subnetworks.Cell. 2014; 158: 808-821Abstract Full Text Full Text PDF PubMed Scopus (157) Google Scholar). The fraction of thin spikes in VAL or VPL was low, and the probability of finding narrow spiking neurons increased from ∼0 to 1 along the electrode tracks into RT (Figures 2C and 2D). We estimated the VAL-VPL and RT boundary at half-max of the probability finding narrow spiking neurons. This yielded an error of 88.5 ± 81.1 μm (VAL) and 50.6 ± 28.9 μm (VPL) from the anatomical boundary, indicating that our data collection and analyses pipeline produced a consistent estimation of recording locations (Figure 2E). With the above pipeline, we collected spikes from 19,690 recording sites across 548 sessions from 43 mice. The recording sites covered 47 thalamic nuclei (94% of total thalamic nuclei; see Table S1 for a full list of thalamic nuclei; Figure 2F). From the recording sites, we sorted 13,441 single units, including 11,836 putative thalamic units and 847 putative inhibitory units (see STAR Methods; Figures S2C and S2D). We analyzed FCC using both multi-unit activity (MUA) and single-unit activity (SUA). We focused on MUA when analyzing population activity, because there were more recording sites, and MUA analyses produced similar results (Figures 3, 4A–4D, 5, 7, S3A–S3I, S3K–S3M, S4B, S5, and S7), and we checked SUA when concerning FCC for individual thalamic neurons (Figures 4E, 6, S3J, S3N, S4C–S4J, and S6). The baseline activity had a wide distribution, with the ventral, geniculate, lateral, and anterior groups of nuclei and RT (e.g., VAL, VPL, LGd-sh, POL, LD, IGL, and RT; SUA > 9.8 spikes/s) having higher activity compared with nuclei adjacent to the midline (e.g., SPFm, SPA, Xi, and PVT; SUA < 5.2 spikes/s; Figures S2E and S2F).Figure 4Topographical Organization of Direct FCC to RTShow full caption(A) Spatial distribution of RT recording sites inhibited by inactivation of indicated cortical area (MOs n = 452; MOp n = 363; SSp n = 511; RSPd n = 179; PTLp n = 235; VIS n = 267; including overlapping sites). Dots, RT recording sites.(B) Hierarchical clustering of RT response patterns following cortical inactivation yielded 6 clusters (Ward’s linkage criterion). Cluster 1, n = 86; cluster 2, n = 156; cluster 3, n = 87; cluster 4, n = 178; cluster 5, n = 111; cluster 6, n = 94; total, 712.(C) Spatial distribution of clusters in RT (left, same orientation as in A) and distribution of each cluster (right). Dots indicate RT recording sites.(D) Left: normalized PSTH (to the baseline 100 ms before laser onset) of each cluster during control (gray) and inactivation (colored line) conditions. Orange bar: duration of laser illumination. Dashed lines indicate laser onset. Right: pattern of cortical inputs to each cluster. Shading indicates SEM, bootstrap across recording sites of each cluster.(E) Example single units in RT. In each panel, the left schematic shows cortical locations providing functional inputs to the unit. The right figures show normalized PSTH of the unit following inactivation (each red line for one cortical location indicated in the left). Black lines indicate control without inactivation. Dashed lines indicate laser onset.See also Figure S4.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 5Topographical Organization of Direct FCCShow full caption(A) Spatial correlation between coordinates of recording sites in each thalamic nucleus and their cortical inputs. Gray dots indicate thalamic nuclei with significant correlation (Pearson’s correlation, p < 0.01). Open circles indicate not significant correlation. Black bar indicates mean.(B) Top: the number of cortical inputs (0–6, MOs, MOp, SSp, RSPd, PTLp, VIS) for each thalamic nuclei. Bottom: rank of the volume of thalamic nuclei. Gray bars indicate nuclei having significant spatial correlation as in (A).(C) Hierarchical clustering of VPM response patterns following cortical inactivation yielded 5 clusters (Ward’s linkage criterion).(D) Spatial distribution of the recording sites in example clusters in VPM (see Figure S5 for other clusters).(E) Clusters in VPM showed distinct dynamics. Top: cortical inactivation locations. Middle (left two panels): normalized population activity of neurons in cluster 1 during control (gray) and inactivation (colored) conditions. Shading indicates SEM. Smoothing window, 20 ms. Middle right: cortical innervation pattern to neurons in cluster 1. Bottom: same as the middle figures but for neurons in cluster 2. Shading, SEM, bootstrap across recording sites of each cluster.(F–H) Example clusters in PO. Same format as in (C)–(E).(I–K) Example clusters