Research Articles, Development/Plasticity/Repair

Differential Recovery of Submodality Touch Neurons and Interareal Communication in Sensory Input-Deprived Area 3b and S2 Cortices

Ruiqi Wu, Pai-Feng Yang, Feng Wang, Qing Liu, John C. Gore and Li Min Chen
Journal of Neuroscience 14 December 2022, 42 (50) 9330-9342; https://doi.org/10.1523/JNEUROSCI.0034-22.2022

Abstract

Cortical reactivation and regain of interareal functional connections have been linked to the recovery of hand grasping behavior after loss of sensory inputs in primates. We investigated contributions of neurons in two hierarchically organized somatosensory areas, 3b and S2, by characterizing local field potential (LFP) and multiunit spiking activity in five states (rest, stimulus-on, sustained, stimulus-off, and induced) and interareal communication after grasping behavior of dorsal column lesioned male squirrel monkeys had mostly recovered. Compared with normal cortex, fMRI, LFP, and spiking response magnitudes to step indentations were significantly weaker. The sustained component of the spiking recovered much better than the stimulus-off response. Correlation between overall spiking and γ LFP remained strong within each recovered areas 3b and S2. The interareal correlations of γ LFP were severely disrupted, except in the resting and stimulus-on periods. Interareal correlation of spiking was disrupted in the stimulus-off period only. In summary, submodality of low threshold mechanoreceptive neurons recovered differentially in input-deprived area 3b and S2 when impaired global hand grasping behavior returned. Slow-adapting-like neurons recovered, whereas rapid-adapting-like neurons did not. Interareal communications were also severely compromised. We propose that slow-adapting-like neurons and afferents in recovered area 3b and S2 mediate recovery of impaired grasping behavior after dorsal column tract lesion.

SIGNIFICANCE STATEMENT Sensory feedback is essential for execution of hand grasping behavior in primates. Reactivations of somatosensory cortices have been attributed to recovery of such behavior after loss of sensory inputs via largely unknown mechanisms. In input-deprived area 3b and S2 cortex, after hand grasping behavior mostly recovered, we found slow-adapting-like neurons were greatly recovered, whereas rapid-adapting-like neurons did not. Communications between area 3b and S2 neurons were severely compromised. We suggest that recovery of slow-adapting-like neurons in input-deprived area 3b and S2 may mediate the recovery of hand grasping behavior.

Introduction

Skilled hand grasping behavior is a hallmark of primate daily life. Sensory feedback is essential for execution of skilled hand grasping behavior. In patients with spinal cord injury, preservation of sensory function is linked to better recovery (Jurkiewicz et al., 2007; Freund et al., 2013; Krishna et al., 2014). We and others have also linked dynamic cortical reactivations and somatotopic and circuitry reorganization in multiple areas (including area 3b, area 1, and S2) to recovery of impaired hand grasping behavior (Florence et al., 1997; Jain et al., 1997; Wang et al., 2013; Yang et al., 2014; Wu et al., 2016, 2017; Fisher et al., 2020; Qi et al., 2021) in a nonhuman primate spinal cord injury model, in which ascending sensory inputs were disrupted by a unilateral dorsal column tract lesion (DCL). In these monkeys, the primary somatosensory cortex S1 area 3b neurons' responses to vibration stimuli at >15 Hz frequency were severely compromised, suggesting that high-frequency vibrotactile information carrying neurons were more affected than low-frequency pressure encoding neurons. S2 neurons' frequency responses, however, were less affected (Wang et al., 2013). We also observed significantly weakened functional connectivity between area 3b and S2 neurons, indicating compromised interareal communication in these two areas (Wu et al., 2017). These previous observations motivated us to further study the status of tactile neurons' adaption properties and their ability to entrain activity in sensory-deprived but recovered areas 3b and S2.

Peripherally, submodality slow-adapting (SA) and rapid-adapting (RA) afferents are segregated and involved in distinct functions (Handler and Ginty, 2021). SA afferents are thought to mediate form and coarse texture perception, whereas that carried by RA afferents are thought to underlie touch frequency perception, motion detection, and fine texture encoding (Blake et al., 1997; Weber et al., 2013; Long et al., 2022). New evidence from multiple mammalian species, including monkeys and humans, suggests that the main organizational principle in the dorsal column is modality: cutaneous fibers tend to cluster medially while deep neuron fibers ascend laterally (Niu et al., 2013). Convergence of submodality responses occurs early in the somatosensory processing in monkeys (Pei et al., 2009). For example, integration has been reported in the cuneate nucleus (Suresh et al., 2021), area 3b neurons exhibited both sustained SA-like and transient RA-like responses (Pei et al., 2009; Carter et al., 2014).

We simultaneously recorded neurons in these two areas and analyzed local field potential (LFP) and spiking activities in different states (resting, stimulation, and induced) because they are linked to distinct brain functions (Lee and Dan, 2012; McCormick et al., 2015). Analyses of LFP and spiking activity in different areas and in different states could provide clues about their potential contributions to the recovery of behavior because these two different signals reflect different aspects of neuronal processing (Reed et al., 2008; Katzner et al., 2009; Hwang and Andersen, 2011; Buzsáki et al., 2012). Specifically, in this study, we (1) used an adaptation index (AI) measure (Pei et al., 2009; Saal and Bensmaia, 2014) to characterize neuron adaption properties in recovered areas 3b and S2; (2) quantified (a) the effects of sensory deprivation on magnitude and duration of fMRI, LFP, and spiking activity during resting, tactile stimulation via probe indentation, and poststimulation-induced states, (b) correlation between spiking and γ-LFP signal, and (c) interarea correlation in each of sensory-deprived and recovered area 3b and S2; and (3) examined how the effects of sensory deprivation on area 3b and S2 neurons differ. We found that sensory loss severely weakened cortical activity in both area 3b and S2 and compromised interareal communication.

Materials and Methods

Animal preparation and surgery

Six adult male squirrel monkeys (Saimiri sciureus) are included in this study. A unilateral DCL between spinal cord cervical segments C4-C5 was conducted on three of them (SM-B, SM-G, and SM-S). fMRI experiments were performed on these 3 monkeys before and months after DCL. Electrophysiology data were acquired from 3 DCL monkeys and 3 normal monkeys (SM-C, SM-K, and SM-4043). This study is a continuation of our previously reported research (Wu et al., 2017). Additional new data collected from the same group monkeys were analyzed and reported here.

Animals were preanesthetized with ketamine hydrochloride (10 mg/kg)/atropine (0.05 mg/kg), and then anesthetized with 0.5%-1.2% of isoflurane delivered with 70/30 N2O/O2 gas mixture to maintain a stable physiological condition for both MRI scans and electrophysiological experiments. Although the actual level may vary across experiments, we typically maintained light anesthesia around 0.7%-0.8% during our fMRI and electrophysiology data acquisitions. The anesthetized animals were intubated and artificially ventilated. After intubation, the animal was placed with its head secured using ear bars and an eye bar. Lactated Ringer's solution was infused intravenously (2-3 ml/h/kg) to prevent dehydration during the course of the study. Arterial blood oxygen saturation and heart rate (Nonin), electrocardiogram, end-tidal CO2 (22-26 mmHg; Surgivet), and respiration (SA Instruments) were externally monitored and maintained. Body temperature was monitored (SA Instruments) and maintained between 37.5°C and 38.5°C via a circulating water blanket (Gaymar Industries). Real-time monitoring was maintained from the time of induction of anesthesia until full recovery. Detailed procedures were similar to those previously reported (Chen et al., 2001, 2002; Reed et al., 2008). All procedures were conducted in accordance with National Institutes of Health guidelines and approved by Animal Care and Use Committees at Vanderbilt University.

Stimulation protocol

Animals' fingers were secured by gluing small pegs to the fingernails and fixing these pegs firmly in plasticine, leaving the glabrous surfaces available for vibrotactile stimulation. The vibrotactile stimulation on an individual digit was generated by a round plastic probe (2 mm in diameter) controlled by a piezoelectric device (Noliac). The piezos were driven by Grass S48 square wave stimulators (Grass-Telefactor). The probe was in light contact with the skin before the vibrotactile stimuli were delivered. Indentation depth of the probe was ∼0.2 mm. For fMRI data acquisitions, vertical indentations of a probe at 8 Hz rate (with 20 ms pulse duration) were presented as 30 s on/off blocks. Seven epochs were typically presented within one imaging run. Typically, 3-5 runs were collected in each fMRI session (day).

In electrophysiological recording sessions, vertical indentations of a probe (in 0.48 mm displacement) were presented on a distal pad for 0.5 s followed by a 2.0 s baseline period and repeated 100 times in each run. Typically, in each session, 3 or 4 runs were collected. We consider the 1 s prestimulus period as the resting-state section, the first 0.1 s period of stimulus presentation as the stimulus onset section, the following 0.4 s stimulus presentation period as the stimulus sustained section, the 0.1 s poststimulus period as the stimulus-off section, and the following 0.4 s poststimulus period as the induced state section (Fig. 1).

Figure 1.
Figure 1.

Diagram for identifying fMRI activation foci and stimulus protocol for electrophysiological recording. A, The locations of area 3b and projected S2 on the squirrel monkey brain surface. Orange square represents the FOV of coronal images. Red square represents the FOV of oblique images. Green square represents the FOV of blood vessel image. B, The coronal T2-weight structural image shows the placements of oblique imaging slices of fMRI (solid red rectangle outlines the slice 1 covering area 3b; dashed red rectangle outlines the slice 3 covering S2). Electrodes in area 3b and S2 are placed perpendicularly to oblique images and recording targets. C, Methods for identifying area 3b and S2 recording sites. The paired area 3b-S2 recording sites are responsive to the same digit. The robust electrophysiology response sites within the fMRI activation region to the corresponding digit were chosen as recording sites. Black arrows indicate recording sites in area 3b (700-900 µm in depth) and S2 (3500-5000 µm in depth). Electrode penetration sites (small dots) on the corresponding blood vessel map represent the response sites to different digits. Green dots represent D1 response sites; cyan dots represent D2 response sites; purple dots represent D3 sites; orange dots represent D4 response sites; navy dots represent D5 response sites. A, Anterior; L, lateral; M, medial; P, posterior; D, dorsal; V, ventral; L, left; R, right. D, Stimulus protocol of electrophysiological experiment. Each typical run includes 100 repetitions of 0.5 s on and 2 s off vibrotactile stimulation on individual digit. We consider the 1 s prestimulating period as “resting” section, the first 0.1 s stimulating period as “on” section, the 0.4 s post “on” section as “sustained” section, the 0.1 s poststimulating period as “off” section, and the 0.4 s post “off” section as “induced” section. Detailed data sampling information is provided in Extended Data Figure 1-1.

Figure 1-1

Run numbers of fMRI and Electrophysiological studies for each animal. Download Figure 1-1, DOCX file.

DCL and behavioral assessment

Unilateral DCL of the cervical spinal cord was performed on the dominant hand side under surgical-level isoflurane (1%-3%) anesthesia. Under aseptic conditions, a portion of the cervical spinal cord was exposed, and the dorsal column tract was sectioned on one side with a fine pair of surgical scissors at caudal C4 level. The dura was replaced with Gelfilm and covered with Gelfoam. Analgesics were administered postsurgically and standard postsurgery care was performed. After DCL, each animal's hand use behavior was evaluated and documented from two aspects: home cage activity and food reaching-grasping-retrieving behavioral testing. For qualitative assessment of home cage hand use, visual observations were used to determine the impairment and recovery of the uses of affected hands. Specifically, immediately after recovery from the surgery, the hand use was carefully monitored and recorded 3 times a week for the first 2 weeks, and then 2 times a week until the behavior returned to normal. We recorded the posture of the hand during the food reaching and retrieving action, frequency of affected hand use during 20 min observation periods, and whether the affected hand could pick up food pellets from a food bin. After a unilateral DCL, all animals showed hand use impairments, which are characterized by unwillingness to use injured hand for cage climbing, abnormal hand posture on holding food, and more false positive food retrieving trials. In the first 2-3 weeks after DCL, all 3 monkeys could not successfully perform food reaching and retrieving tasks (for detailed descriptions, see Qi et al., 2013), only these qualitative visual observations were used to determine the recovery of the affected hand. Using these criteria, lesioned subjects all exhibited full recovery by the time of fMRI and electrophysiology recording. The details of the behavioral assessment and surgical procedures can be found in previous publications (Qi et al., 2011, 2013; Chen et al., 2012; Yang et al., 2014; Wu et al., 2016).

MRI data acquisition and analysis

Each monkey underwent multiple fMRI sessions before and after DCL. MRI scans were performed on a 9.4 T Inova magnet (Varian Medical Systems), using a 3-cm-diameter surface transmit-receive coil centered over area 3b and S2 contralateral to the stimulated hand. Four 2-mm-thick oblique image slices were centered over the central and lateral sulci to maximize signal noise ratio. Usually, the first slice covered SI and the third slice covered S2 (Fig. 1A,B). fMRI data were acquired using a gradient EPI sequence (TR = 1500 ms, TE = 16 ms, and resolution = 0.55 × 0.55 × 2 mm3). Each fMRI run has 300 image volumes. fMRI data were then analyzed and spatially interpolated to overlay on T2-weighted gradient echo high-resolution anatomic images (TR = 200 ms, TE = 16 ms, and resolution = 0.078 × 0.078 × 2 mm3) for display. After baseline fMRI data collection, monkeys received unilateral DCL at cervical (caudal C4) spinal cord level. fMRI data acquisition was repeated after DCL. The last post-DCL fMRI scan, which typically was ∼1-2 weeks before the electrophysiology recording session, was used in the current study to map the digit activation. Both surface and trans-cortical blood vessel features were easily identifiable on T2-weighted images and were used for coregistration of MRI maps acquired in different image sessions, and to the optical surface blood vessel maps obtained later in the microelectrode mapping sessions (see Fig. 2C).

EPI data were preprocessed with physiological motion correction (3dretroicor, AFNI), slice timing, and head motion correction (3dTshift, AFNI), and spatial smoothing (Gaussian kernel FWHM = 1 mm, 3dmerge, AFNI). The stimulus-evoked EPI data were temporally smoothed with a low-pass filter with a 0.25 Hz cutoff frequency (3dFourier, AFNI). Activation maps were created by voxel-wise correlation analysis, using the HRF convolved stimulus presentation paradigm (3dDeconvolve, AFNI). Activation maps were registered on corresponding T2-weighted anatomic images using a linear image registration tool (flirt, FSL), and then displayed as statistical t value maps. Activation maps were generated from t value maps (with a threshold of t value = 2, p < 0.05) of all the runs within the same session day (see Fig. 2A,B). We quantified the percentage of detected activation for area 3b and S2 before and after DCL. BOLD signal time courses were extracted from one single peak voxel that showed the highest probability of detecting activation across multiple imaging runs in each area (area 3b and S2) to quantify amplitudes of BOLD responses to stimuli. The peak responses in both area 3b and S2 were compared before and after DCL.

Identification of area 3b and S2 neurons

In each animal, robust fMRI activations were detected in area 3b and S2 regions. fMRI activation maps were overlaid on brain's surface blood vessel maps that served as landmarks for localizing fMRI activation foci and mapping spatial shift of activation foci after DCL. We then use these activation maps to guide the placements of dense microelectrode penetrations around the central sulcus and into the upper bank of the lateral sulcus where digits representations in area 3b and S2 reside (Fig. 1B,C). Under isoflurane anesthesia, a craniotomy was made over the somatosensory cortex, and the dura was removed for mapping and recording. Single microelectrodes (FHC) that were epoxylite-coated tungsten with exposed standard sharp tip (<3 μm) of ∼1 mΩ impedance (measured at 1 kHz) penetrated through cortical layers of area 3b and then further advanced into the upper bank of the lateral sulcus where the S2 region is located. At each electrode penetration site, the microelectrode was advanced in 300 µm increments, and its tip depth was tracked and logged. Preferred stimulus type (i.e., stroke, squeeze, light touch), response strength (scored in six levels), and receptive field properties were characterized. Receptive fields of neurons were identified by palpating skin areas on the contralateral hand to the recording side while listening to the audio amplifier for spike activity and viewing traces of action potentials on a display. After the completion of the mapping, we were able to identify area 3b, area1, and S2 and area borders based on the characteristics of receptive field properties, preferred stimuli, and somatotopic organization of the digits (Kaas et al., 1984; Jain et al., 2008; Wang et al., 2013). We chose area 3b and S2 neurons for recording according to mapping results. To examine how area 3b and S2 neurons communicate, we stimulated the finger that could simultaneously elicit electrical activity for both area 3b and S2 neurons.

Electrophysiological data acquisition, preprocessing, and analysis

Data acquisition

Electrophysiological recordings were performed under light anesthesia that was similar to that for fMRI data acquisition. When recording, the exposed cortex was covered with 4% agar mixed with Ringer's solution to provide stability. Pairs of electrodes were placed in recognized responsive regions in area 3b and S2. Spiking and LFP signals were then recorded by using a Multichannel Acquisition Processor System (Plexon). Electrophysiological signals were passed through a unit-gain head-stage and a preamplifier and then fed through two separate analog filters: one for higher frequencies of neuronal spikes (100-8000 Hz) and one for lower-frequency LFP signals (0.7-300 Hz). The spiking signals were then amplified and digitized at 30 kHz sampling rate. Single units were sorted with offline sorter software (Plexon). LFPs were sampled at 1 kHz.

Signal preprocessing and analysis

We used custom-written codes in MATLAB (The MathWorks, RRID:SCR_001622) for the electrophysiological data processing and statistical analysis, after spike sorting (Plexon) and peristimulus time histogram (PSTH) analysis (NeuroExplorer, Nex). Raw LFP data were filtered by a 60 and 120 Hz band stop chebyshef filter to remove the power line frequency interference, and then were bandpass filtered between 1 and 150 Hz for quantification. Evoked LFP signals were averaged across 100 trials within runs for presentation. The power spectral density of 1 Hz resolution was calculated, and the LFP power was transformed into dB scale (10 × log10) to build a time-frequency graph. Different LFP frequency bands presumably represent different functional properties of neurophysiology, so LFP signals were filtered as δ band (1-4 Hz), θ band (5-8 Hz), α band (9-12 Hz), β band (13-30 Hz), low γ band (l_γ, 30-50 Hz), high γ band (h_γ, 50-100 Hz), and very high γ band (vl_γ, 100-150 Hz) for further analysis. The power of different bands at resting state (1 s prestimulus section), stimulating state (including 0.1 s onset stimulus section, 0.4 s sustained stimulus section, and 0.1 s offset stimulus section), and induced state (0.4 s post-offset section) were calculated, and then were rescaled by min-max normalization with the formula of S′ = (S – min(S))/(max(S) – min(S)). Five sections were divided according to response feature of LFP and spiking signals (see Figs. 3, 6A,B). Each temporal period represents a different brain functional state. Power change percentages of all frequency bands at stimulating state and induced state were calculated by subtracting and then dividing their corresponding baseline signals at resting state.

Multiunit spikes were also analyzed. To extract spikes and remove transient noise and artifacts, single-unit spiking activity was first isolated, and then combined into multiunit activity for further analysis with Rasputin software (Plexon). PSTH of spiking activity of area 3b and S2 neurons was calculated and plotted in different temporal windows (states): before (baseline), during (stimulating), and after (induced) each vibrotactile stimulus event. We used 10 ms as the bin width (Reed et al., 2008; Wang et al., 2013). PSTHs from 100 trials were averaged within each run, and then averaged across runs. Spiking counts at the same state were averaged at the group level for comparison between states and between normal and lesioned conditions. Percentages of spiking rate changes at stimulating state (including stimulus-on, sustained, and off periods) and induced state were calculated by subtracting and then dividing their corresponding baseline spiking rate before stimulus onset. Activity duration was estimated by the formula of t0.5/ln2 ∼= 1.443 * t(0.5), usually used to calculate mean lifetime in decaying quantification (Pomme, 2015), where t(0.5) is the time from peak activity to half-peak activity. AI was calculated by AI = (2/π)*tan−1(Roff/Rs), where Roff is the spiking rate at offset section and Rs is the spiking rate at sustained section (Pei et al., 2009).

Power of LFP γ band (50-150Hz) of area 3b and S2 was extracted, and then averaged within each run. Temporal resolution of γ band power was downsampled to 10 ms (i.e., points within every 10 ms were binned and averaged as one data point), generating a temporal resolution the same as PSTH of spiking activity.

Correlation analysis between signals in different states within and between cortical areas

Pearson correlations between time courses of γ band LFP power and firing rates at different states (including resting, stimulus-on, sustained, stimulus-off, and induced state) were calculated in three conditions: between γ band LFP and spiking within each area 3b and S2, γ band LFP power between area 3b and S2, and spiking activity between area 3b and S2. Correlation values were compared in normal versus sensory input-deprived cortex.

Statistical analysis

For fMRI studies, 12 runs of the prelesion condition and 12 runs of the postlesion condition from 3 monkeys (SM-G: four runs; SM-B: 5 runs; SM-S: three runs) were collected. A total of 84 stimulus epochs were acquired in each of the normal and DCL groups. The time courses of 7 epochs of each run were averaged, and peaks of signal changes were extracted. For group analysis, fMRI signal changes of all runs within individual animals were averaged, and then signals were compared between prelesion and postlesion conditions across animals. Electrophysiological data, including 14 runs from normal animals (SM-C: four runs; SM-K: 6 runs; SM-4043: four runs) and 10 runs from lesioned animals (SM-G: three runs; SM-B: three runs; SM-S: four runs), were analyzed. Detailed data sampling information is provided in Extended Data Figure 1-1. The signals of 100 trials within each run were averaged, and then signals were averaged within each individual animal. Group statistical analysis was performed across animals between normal and lesion groups. Normalized LFP power, LFP signal change, spiking counts, spiking rate change, and spiking duration, and correlation between LFP γ band and spike signal of area 3b and S2 were calculated for statistical analysis. We performed nonparametric permutation tests for two-group comparisons and nonparametric Kruskal–Wallis H tests for multiple-group comparisons. Statistical analyses of Kruskal–Wallis H tests were performed in SPSS (IBM, RRID:SCR_002865), and permutation tests were performed using MATLAB (https://github.com/lrkrol/permutationTest). p < 0.05 was considered statistically significant. Results are presented as mean ± SE. For correlation analysis, we calculated both r value and p value (corrcoef, MATLAB). p < 0.05 (and a confidence level of >95%) is considered as statistically significant.

Results

Weakened and location shifted fMRI responses to tactile stimulation in sensory input-deprived area 3b and S2

We compared tactile stimulus-evoked fMRI signals obtained before and after DCL in both areas. Prelesion area 3b and S2 showed robust stimulus-evoked fMRI activation (Fig. 2A, maximum t ∼= 5). However, the fMRI activation to the same stimuli was drastically weakened months after sensory deprivation (Fig. 2B, maximum t ∼= 3, presented with the same threshold of Fig. 2A). Centers of activation in both areas shifted laterally and posteriorly (Fig. 2C). The fMRI signal changes in area 3b and S2 decreased in all 3 animals after DCL (Fig. 2D). Taking SM-G, for example (Fig. 2D, left column group), BOLD signal changes significantly from 1.83 ± 0.63% to 0.51 ± 0.15% in area 3b (permutation test, p = 0.01; ESCohen's d = 1.44; 95% CI of the Hodges–Lehman median difference = (0.3%, 3.4%); n = 4 in both normal and lesioned conditions) and from 1.26 ± 0.11% to 0.68 ± 0.12% in S2 (permutation test, p = 0.03; ESCohen's d = 2.48; 95% CI of the Hodges–Lehman median difference = (0.2%, 0.9%); n = 4 in both normal and lesioned conditions). There was no significant statistical difference between fMRI signal changes in prelesion and postlesion groups, although signal decreases were observed in all 3 animals (Fig. 2D, right column group). These results extracted from activation maps were consistent with our previous findings extracted from activation probability maps (Wu et al., 2017).

Figure 2.
Figure 2.

Tactile stimulus-evoked fMRI activations in area 3b and S2 before and after DCL and illustration of paired area 3b-S2 recording sites in postlesion cortex. A, B, fMRI activation maps to tactile stimulation before (A) and after DCL (B) in area 3b (left, slice 1) and S2 (right, slice 3) in Monkey SM-B. Color scale represents t values. C, Comparison of D2 and D3 stimulation-evoked fMRI activation foci before and after lesion in both area 3b and S2. S2 activations were projected to brain surface and overlayed with area 3b activations for comparison. Blue shading represents area 3b activation pattern outline. Green shading represents S2 activation pattern outline. Dark color shading represents outline of the prelesion activation. Light shading represents outline of the postlesion activation. Locations of activation foci shifted in both areas. Brain surface blood vessel map is used as landmarks to guide microelectrode recording in each animal. Electrode penetration sites (dots) on vessel map show the response sites to different digits. Green dots represent D1 response sites; cyan dots represent D2 response sites; purple dots represent D3 sites; orange dots represent D4 response sites; navy dots represent D5 response sites. A, Anterior; L, lateral; M, medial; P, posterior. D, Percentage fMRI signal changes in area 3b and S2 before (purple) and after DCL (gray) within (the left three column groups) and across animals (the right column group). N = 4 in SM-G; N = 5 in SM-B; N = 3 in SM-S; N = 3 in normal and lesioned groups. Permutation tests. *p < 0.05. Error bar indicates SE.

Weakened LFP signal amplitude and power in sensory input-deprived area 3b and S2

Figure 3 showed examples of event-averaged LFP amplitude and power spectrum plots in 3 normal (Fig. 3A1–C2) and 3 lesioned animals (Fig. 3D1–F2). In the normal cortex, a repeated vertical indentation of a distal finger pad elicited robust stimulus-phase locked LFP amplitude changes (Fig. 3A1,B1,C1) and power increases across the entire LFP frequency band (Fig. 3A2,B2,C2) at both stimulus-on and stimulus-off periods in area 3b and S2. In contrast, the LFP signal amplitude (Fig. 3D1,E1,F1) and power increases (Fig. 3D2,E2,F2) were much weaker during the stimulus-on period and barely detectable during the stimulus-off period in the sensory input-deprived cortex.

Figure 3.
Figure 3.

Examples of event-averaged LFP amplitudes and power spectrums in area 3b and S2 in 3 normal and 3 DCL monkeys. A1, B1, C1, Averaged stimulus-evoked LFP in area 3b (area 3b) (left) and S2 (right) in normal monkey SM-K (A1), SM-C (B1), and SM-4043 (C1). Black line next to x axis indicates the stimulating periods. A2, B2, C2, Power spectrums of area 3b (left) and S2 (right) in normal monkey SM-K (A2), SM-C (B2), and SM-4043 (C2). D1, E1, F1, Averaged stimulus-evoked LFP of area 3b (left) and S2 (right) in lesioned monkey SM-G (D1), SM-B (E1), and SM-S (F1). D2, E2, F2, Power spectrums of area 3b (left) and S2 (right) in lesioned monkey SM-G (D2), SM-B (E2), and SM-S (F2).

We next quantified the normalized power differences at seven LFP frequency bands (δ, θ, α, β, low-γ, high-γ, and very high-γ) and at five states. In the normal cortex, robust power increases during stimulus-on and stimulus-off periods were present at high-frequency bands in both area 3b and S2 (comparing red and purple columns with green columns in Fig. 4A,C; Extended Data Fig. 4-1A,C, Extended Data Fig. 4-2A,C). The power increases were statistically significant at the very high γ band for on and off periods of tactile stimulation in both area 3b (nonparametric Kruskal–Wallis H tests, χ2 = 13.38, p = 0.01, total n = 15, degrees of freedom = 4; for on period: p = 0.001, ESepsilon-squared = 0.85, 95% CI of the Hodges–Lehman median difference = (0.004, 0.015); for off period: p = 0.015, ESepsilon-squared = 0.63, 95% CI of the Hodges–Lehman median difference = (0.001, 0.002)) and S2 (nonparametric Kruskal–Wallis H tests, χ2 = 12.27, p = 0.015, total n = 15, degrees of freedom = 4; for on period: p = 0.003, ESepsilon-squared = 0.76, 95% CI of the Hodges–Lehman median difference = (0.002, 0.019); for off period: p = 0.022, ESepsilon-squared = 0.60, 95% CI of the Hodges–Lehman median difference = (0.0003, 0.004)). The overall power increases were greater in S2 than those in area 3b (compare the last column groups in Fig. 4A,C). In the sensory input-deprived cortex, two features were noticeable. First, the overall powers in resting baseline and induced periods were higher than those in the normal cortex (compare green and brown columns in Fig. 4A,C with Fig. 4B,D; and Extended Data Fig. 4-1A,C with Extended Data Fig. 4-1B,D). Second, the power differences in the five states became much smaller (Fig. 4B,D; Extended Data Fig. 4-2B,D). These changes occurred in both areas 3b and S2.

Figure 4.
Figure 4.

LFP powers in normal and input-deprived area 3b and S2 cortex at five different states. A, B, The normalized power of area 3b in normal (A) and lesioned animals (B) at different LFP frequency bands. C, D, The normalized power of S2 in normal (C) and lesioned animals (D) at different LFP frequency bands. Nonparametric Kruskal–Wallis H tests. n = 3 in both normal and lesioned groups. *p < 0.05. Error bar indicates SE. L, Low; H, high; VH, very high. LFP powers at low-frequency bands in normal and input-deprived area 3b and S2 cortex at five different states are presented in Extended Data Figure 4-1. Profiles of LFP γ band (50-150 Hz) powers in normal and input-deprived area 3b and S2 cortex at five different states across animals are shown in Extended Data Figure 4-2.

Figure 4-1

LFP powers at low frequency bands in normal and input-deprived area 3b and S2 cortex at five different states. (A-B) The normalized power of area 3b in normal (A) and lesioned animals (B) at delta, theta, and alpha bands. (C-D) The normalized power of S2 in normal (C) and lesioned animals (D) at delta, theta, and alpha bands. N = 3 in both normal and lesioned groups. Error bar indicates SE. Download Figure 4-1, TIF file.

Figure 4-2

Profiles of LFP Gamma band (50Hz-150Hz) powers in normal and input-deprived area 3b and S2 cortex at five different states across animals. (A-B) The normalized power of area 3b in each normal (A) and lesioned animals (B) at gamma band. (C-D) The normalized power of S2 in each normal (C) and lesioned animals (D) at gamma band. N = 3 in both normal and lesioned groups. Nonparametric Kruskal–Wallis tests, *p < 0.05, Error bar indicates SE. Download Figure 4-2, TIF file.

Percentage changes of LFP power responses at different frequency bands in normal and input-deprived cortices

To understand the degree of changes in LFP power after sensory deprivation, we normalized and compared the LFP signal power changes in area 3b (Fig. 5A) and S2 (Fig. 5B) at stimulation on, sustained, stimulus-off, and induced states, using the resting-state signals as baseline. Regardless of area 3b or S2, the significant LFP signal rises were observed in almost all states in normal cortex (see green statistical stars under purple columns). After DCL, the stimulation-evoked LFP signal decreased across all frequency bands in both area 3b and S2 (compare the purple and gray columns in Fig. 5A and Fig. 5B, respectively). Stimulus-induced LFP responses were significantly decreased at stimulation on and sustained states in all frequency bands, except the very high γ band, in area 3b (compare purple and gray columns in the first two column groups in Fig. 5A). Significant stimulus-induced LFP signal decreases were also present at off state in high-frequency bands (including β, low γ, high γ, and very high γ bands) in area 3b (compare purple and gray columns in the third column groups in Fig. 5A). However, the LFP signal reductions were only significant at off state in all frequency bands in S2 (Fig. 5B, third column groups).

Figure 5.
Figure 5.

Normalized LFP signal changes at different frequency bands at different states in normal versus input-deprived conditions. LFP signal changes in area 3b (A) and S2 (B) in normal (purple) and DCL (gray) monkeys for stimulus states (on, s, off, and i) normalized by resting state (r). Green stars indicate statistical differences between signal change and zero. Permutation tests, n = 3 in both normal and lesioned groups. *p < 0.05. Error bar indicates SE.

Effects of DCL on temporal dynamics of multiunit spike activity and LFP signals and changes in adaptation

Figure 6 shows the differences in the group-averaged dynamic profiles of spiking (PSTH) (Fig. 6A) and γ LFP power (Fig. 6B) before, during, and after a sustained 500 ms probe indentation in normal (purple) and lesioned (gray) cortex. In normal cortex, the firing rates of neurons increased drastically at the stimulus-on as well as at the stimulus-off in both areas. There was sustained firing activity during the consistent probe indentation period in area 3b, but there was little in S2 cortex. Firing activity maintained at a moderate level during the poststimulus-induced activity periods (Fig. 6A, purple). In the input-deprived cortices, the stimulus-evoked spiking rate increases in area 3b were significantly reduced at the stimulus-on period (last ∼30 ms; Fig. 6A, gray plot) and exhibited almost no activity at the stimulus-off period. The firing rates of S2 neurons exhibited very similar trends to that of area 3b neurons in both normal and lesioned conditions (Fig. 6A). However, firing activity of normal S2 neurons lasted much longer during stimulus-on and stimulus-off periods, compared with that of normal area 3b. The temporal profiles of γ band LFP in different states were similar to the observations in firing rates (Fig. 6B). Activities at the baseline, sustained, and induced states were close to none for S2.

Figure 6.
Figure 6.

Effects of sensory deprivation on temporal dynamics of spike rates and LFP γ power in area 3b and S2. A, The spiking counts of area 3b (left) and S2 (right) neurons before, during, and after vertical probe indentation. B, The LFP γ band power of area 3b (left) and S2 (right) neurons in normal (purple) and lesioned (gray) animals. Yellow shading represents on; green represents sustained; blue represents off. C, The proportion of area 3b and S2 neurons in normal and lesioned monkeys with AI ranging 0-1. D, AI of area 3b and S2 neurons in normal (purple) and lesioned (gray) monkey. Permutation tests were used in group analysis. N = 3 in both normal and lesioned groups. *p < 0.05. Error bar indicates SE.

We calculated the AI and evaluated the changes in neurons' adaptation properties after sensory deprivation by plotting the proportion of neurons as a function of AI value (Fig. 6C). A shift from a high number of RA-like neurons in the normal cortex to a high number of SA-like neurons in the lesioned cortex was apparent, indicating that a few RA-like neurons were recovered in both cortical areas. Generally, the AIs of lesioned animals were smaller than those of normal animals in both area 3b (compare the purple columns with gray columns in the left column groups Fig. 6D) and S2 (compare middle column groups Fig. 6D). Group analysis results showed that AI decreased significantly in both input-deprived area 3b and S2 (Fig. 6D, right column groups; for area 3b: permutation test, p = 0.048, ESCohen's d =3.03, 95% CI of the Hodges–Lehman median difference = (0.126, 0.492), n = 3 in both normal and lesioned groups; for S2: permutation test, p = 0.048, ESCohen's d = 2.76, 95% CI of the Hodges–Lehman median difference = (0.102, 0.477), n = 3 in both normal and lesioned groups)). The mean AI of S2 neurons was greater than that of area 3b neurons, but the difference was not statistically significant (Fig. 6D).

Effects of sensory deprivation on spike rate changes

We next quantified the spiking activity and found that sensory deprivation had similar effects on spiking activity as on LFP signals. In the normal cortex, firing rates were significantly increased during stimulus-on and stimulus-off periods in area 3b (Fig. 7A, left group of columns; nonparametric Kruskal–Wallis H tests, χ2 = 11.47, p = 0.022, total n = 15, degrees of freedom = 4; for stimulus-on period: p = 0.006, ESepsilon-squared = 0.71, 95% CI of the Hodges–Lehman median difference = (22.5, 50.6); for off period: p = 0.036, ESepsilon-squared = 0.55, 95% CI of the Hodges–Lehman median difference = (18.7, 33.6)), and during on periods in S2 (Fig. 7A, right group of columns; nonparametric Kruskal–Wallis H tests, χ2 = 10.37, p = 0.035, total n = 15, degrees of freedom = 4; for stimulus-on period: p = 0.018, ESepsilon-squared = 0.62, 95% CI of the Hodges–Lehman median difference = (21.7, 69.8)) compared with baseline firing rates. In the input-deprived cortex, the overall firing rates of area 3b and S2 neurons were much smaller (Fig. 7B). The firing rate differences between the resting and sustained states, and between the resting and stimulation off state diminished, indicating that both area 3b and S2 neurons lost the ability to maintain spiking activity after the stimulus stopped (Fig. 7B). When we normalized the firing rates and directly compared firing rates in each state between normal and lesioned cortex, we found that the firing rates decreased in all four states and only decreased significantly in stimulus-off period in both lesioned area 3b (see the third column group in Fig. 7C; permutation test, p = 0.048, ESCohen's d = 1.71, 95% CI of the Hodges–Lehman median difference = (1.70, 14.4), n = 3 in both normal and lesioned groups) and S2 (see the third column group in Fig. 7D; permutation test, p = 0.048, ESCohen's d = 1.78, 95% CI of the Hodges–Lehman median difference = (0.81, 16.02), n = 3 in both normal and lesioned groups). The spike rate changes in each period, however, were significant in both normal area 3b and S2, except the sustained and induced periods in S2 (Fig. 7C,D, green statistical stars under purple columns). Consistent with the group PSTH plot, stimulus-on- and stimulus-off-locked firing durations were longer in S2 than area 3b in normal cortex. But they were not statistically significant. This difference disappeared in sensory-deprived cortex (Fig. 7E,F). Stimulus-off-locked firing durations were significantly shortened after DCL in both area 3b (compare purple column and gray column in the left column group in Fig. 7F; permutation test, p = 0.048, ESCohen's d = 3.31, 95% CI of the Hodges–Lehman median difference = (6.81, 28.19), n = 3 in both normal and lesioned groups) and S2 (compare purple column and gray column in the right column group in Fig. 7F; permutation test, p = 0.048, ESCohen's d = 3.55, 95% CI of the Hodges–Lehman median difference = (15.65, 44.41), n = 3 in both normal and lesioned groups).

Figure 7.
Figure 7.

Effects of sensory deprivation on spike rate changes of area 3b and S2 neurons during stimulation and induced periods in normal versus DCL monkeys. A, B, Absolute spiking rates of area 3b and S2 neurons at different states in normal (A) and lesioned animals (B). C, D, Percentage of spiking rate changes of area 3b (C) and S2 (D) neurons at different states normalized by resting state (r) for normal (purple) and lesioned (gray) monkeys. Green stars indicate statistical differences between spike rate change and zero. E, F, Activity durations (FWHM) of on-response (E) and off-response (F) periods. Nonparametric Kruskal–Wallis H tests were used in A, B, and permutation tests were used in C–F. *p < 0.05. n = 3 in both normal and lesioned group. Error bar indicates SE.

Correlation of LFP γ band and spiking signals at different states within and between areas

The overall spiking rate and γ power during stimulus sustained, stimulus-off, and induced was highly correlated in both normal area 3b and S2 (Fig. 8A, purple columns). After sensory deprivation, the mean correlations dropped overall. The spike-γ band LFP correlation decreased significantly after DCL at induced states in both area 3b (Fig. 8A; permutation test, p = 0.048, ESCohen's d = 2.37, 95% CI of the Hodges–Lehman median difference = (0.10, 0.58), n = 3 in both normal and lesioned groups) and S2 (Fig. 8A; permutation test, p = 0.048, ESCohen's d = 1.93, 95% CI of the Hodges–Lehman median difference = (0.12, 0.83), n = 3 in both normal and lesioned groups).

Figure 8.
Figure 8.

Correlation of the spiking and γ band signals at different states in normal versus input-deprived area 3b and S2. A, Correlation of spike and LFP γ band signal in area 3b and S2 at different states in normal and lesioned animals. B, Correlation of LFP γ band signals between area 3b and S2 at different states in normal (purple) and lesioned (gray) animals. C, Correlation of the spiking signals between area 3b and S2 at different states in normal (purple) and lesioned (gray) animals. Permutation tests, n = 3 in both normal and lesioned group. *p < 0.05. Brown lines indicate Pearson correlation statistical cutoffs at r = 0.15, 0.2, 0.65, 0.35, 0.65, and 0.35 (from left to right, respectively). p < 0.05. Error bar indicates SE.

Induced responses after stimulation are involved in the interareal communication (Varela et al., 2001; Adjamian, 2014). In our case, changes in area 3b induced period may influence S2 neurons activity. To characterize this potential influence, we plotted correlation at different states between area 3b and S2 (Fig. 8B). In the normal cortex, LFP γ band signals in area 3b and S2 at each state were all highly correlated (Fig. 8B, purple columns). In the sensory-deprived cortex, the overall correlation significantly dropped (compare the left group of purple and gray columns in Fig. 8B; permutation test, p = 0.048, ESCohen's d = 1.66, 95% CI of the Hodges–Lehman median difference = (0.04, 0.40), n = 3 in both normal and lesioned groups). The significantly decreased correlations were present in sustained (permutation test, p = 0.048, ESCohen's d = 4.68, 95% CI of the Hodges–Lehman median difference = (0.17, 0.43), n = 3 in both normal and lesioned groups), stimulus-off (permutation test, p = 0.048, ESCohen's d = 9.80, 95% CI of the Hodges–Lehman median difference = (0.48, 0.73), n = 3 in both normal and lesioned groups), and induced states (permutation test, p = 0.048, ESCohen's d = 4.13, 95% CI of the Hodges–Lehman median difference = (0.40, 0.88), n = 3 in both normal and lesioned groups).

Spiking activity in area 3b and S2 was also highly correlated overall in stimulus-on and stimulus-off states in normal animals, but not at the resting and induced states (Fig. 8C, purple columns). The correlations significantly dropped after sensory deprivation at stimulus-off state (compare the fifth group of purple and gray columns in Fig. 8C; permutation test, p = 0.048, ESCohen's d = 4.41, 95% CI of the Hodges–Lehman median difference = (0.38, 0.96), n = 3 in both normal and lesioned groups). The LFP γ and spiking correlation between area 3b and S2 significantly differed in sustained and induced states (for LFP only).

Discussion

Differential recovery of RA-like versus SA-like neurons in sensory input-deprived area 3b and S2

It is interesting that, even through by the time the impaired food grasping behavior, which is measured by the ability to successfully retrieve small sugar pellets, largely recovered (Qi et al., 2013; Wu et al., 2017), fMRI signal, LFP power, and spiking activity in the recovered area 3b and S2 cortex were still significantly weaker than normal cortex. Neurons had exhibited more sustained SA-like responses to step indentations (Pei et al., 2009; Carter et al., 2014), compared with normal cortex. Specifically, the initial transient RA stimulus-on response and sustained SA activity seemed to have recovered to a greater degree than the stimulus-off response. The loss of transient RA-like responses during stimulus-off periods was pronounced in both cortices. The lost ability of area 3b and S2 neurons to respond to the removal of probe contact indicate that RA afferents likely are more affected by DCL and did not fully recover (Figs. 6 and 7).

The property changes of area 3b and S2 cortical neurons likely resulted from differential recovery of SA-like versus RA-like afferents occurring at the spinal cord, and convergence of submodality neurons at the supraspinal level. For example, a recent paper by Bensmaia's group has shown that SA-like and RA-like submodality convergence was observed in the dorsal column nucleus (cuneate nucleus) (Suresh et al., 2021). It is unclear why SA afferents mostly recovered while RA-like did not months after DCL. One possibility is that a high proportion of SA afferents ascends to supraspinal regions through nondorsal column pathways; therefore, SA afferents were less affected by DCL. Notably, patients who had a focal lesion at the dorsal column pathway primarily exhibited deficits in somatic percepts associated with transient signals carried by RA afferents (Bors, 1979). This observation further supports the possibility that a DCL affected more RA versus SA afferents. The second possibility is that the spared residual dorsal column afferents and their secondary pathways (Qi et al., 2013) somehow were able to compensate the lost functions of SA afferents better than RA afferents. Central factors, such as compromised temporal inhibition in supraspinal regions, including area 3b and S2, could also be a possible cause. The AI mostly reflects the submodality composition of the input. However, it is also somewhat dependent on central factors. For example, the sustained responses of a neuron to its preferred stimulus (step indentation in our case) will be stronger than that to a nonpreferred stimulus (Gao et al., 2016). The greater preponderance of RA-like responses in normal S2 likely reflects the fact that S2 neurons are more selective than S1 neurons; thus, the central contribution to the AI is stronger. Anesthesia is also liable to impact S2 responses more than S1 responses. These possibilities warrant further investigation.

Since gross hand use behavior (e.g., food reaching and grasping behavior) was mostly recovered by the time of data collection, we speculate the involvement of recovered SA afferents and SA-like neurons in area 3b and S2 is adequate for mediating this behavior. The impacts of lost and unrecovered RA neurons could be reflected in other type of fine-controlled hand and digit movement. Supporting this speculation, our most recent video-based analysis using DeepLabCut (Mathis et al., 2018) captured some hand use deficits that were not detected with the scored behavioral assessment system. A future comprehensive behavioral analysis might provide some clues about the roles of RA afferents and neurons in mediating skilled hand use.

Correlation between γ LFP and spiking activity during stimulation and implication for information integration within each area

Synchronized γ LFP and spiking activity have been linked to behavior prediction (Khamechian et al., 2019). Thus, disruption of this synchrony may be associated with or contribute to compromised behavior. In sensory-deprived area 3b and S2 cortex, we observed dissociation between spiking and broadband (Wu et al., 2017) and γ band LFP signals. During information processing and integration, γ LFP signals and spiking activity are considered as cortical inputs and output signals. γ oscillations play an important role in neural processing, such as corticospinal interaction (Schoffelen et al., 2005) and sensorimotor integration (Roelfsema et al., 1997; Womelsdorf and Fries, 2006). Disrupted correlation between spiking-γ oscillations in S2 may influence the response gain of neurons (Azouz and Gray, 2000, 2003; Shu et al., 2003) or control the timing and probability of action potential generation in pyramidal cells (Hasenstaub et al., 2005) in neural processing.

In future studies, delivery of complex haptic stimuli, such as multidigit complex stimuli and use of the advanced machine learning-based DeepLabCut system, are needed for assessing area 3b's and S2's contribution to the global recovery of hand grasping behavior after DCL. The specific contribution of γ and high-γ (50-200 Hz) LFP activities to the behavior recovery (Liu and Newsome, 2006; Womelsdorf et al., 2006; Lee and Lisberger, 2013; Smith et al., 2015; Khamechian et al., 2019), the hand grasping behavior in our case, also requires further investigation.

Loss of activity in induced state and compromised communication between sensory-deprived area 3b and S2 neurons

Entrained neuronal activity during induced periods is linked to interareal communication. Stimulus-induced spiking (Fig. 7C) and LFP response in area 3b (Fig. 5A) are robust in the normal cortex. When the sensory inputs from spinal cord were deprived, these neurons almost completely lost their ability to entrain firing after stimulation.

Our results also revealed a greatly compromised communication between input-deprived area 3b and S2 (Fig. 8B,C). Significantly disrupted interareal correlation during phasic stimulus-off-period suggest the compromised communication may be neuronal submodality (RA) specific. Compromised correlations between γ LFPs also suggest the afferent drives from area 3b to S2 was severely weakened and did not fully recover. Correlation between area 3b and S2 neurons remained during stimulus-on period for both LFP and spikes. Together, the loss of induced response and correlated signal changes between area 3b and S2 neurons indicates that area 3b and S2 neurons' communication ability was severally compromised after driving sensory inputs were disrupted. This finding of compromised communication between area 3b and S2 neurons during processing, along with our previous report of disruption in resting state functional connectivity between these two areas (Wu et al., 2017) underscored the abnormality of interareal connections, during stimulation processing, induced period and resting state.

One final note is that the lower correlations may simply reflect the overall decreased activity in input-deprived cortex. We do not think this is the case because although the overall activity in input-deprived areas is generally weaker, the weakening is not uniformly distributed and primarily occurred in stimulus-off and induced periods as shown in Figures 5–7. To specifically address this concern, we directly compared the correlations between spike and γ LFP activity within each area, and between area 3b and S2 at different states (resting, on, sustained, off, and induced) before and after DCL (Fig. 8). Figure 8B, C shows that both LFP and spiking rate during stimulus-on period remained comparable in normal versus DCL conditions. The drops in correlation values primarily occurred in the stimulus-off and induced periods when the activity was weak as shown in Figure 8. Thus, these observations indicate that reduced correlation was not directly associated with low neuronal activity. Together, these differential reductions in correlations led us to believe that the lower correlations after DCL reflect the weaker interareal communication, not a biased outcome of decreased activity.

In conclusion, in sensory-deprived area 3b and S2 cortex, neuronal spiking and LFP activities at different brain states were greatly weakened even when the monkey's impaired hand grasping behavior was almost recovered. Overall features of changes were similar between area 3b and S2 neurons. SA-like neurons mostly recovered, while RA-like neurons did not. Interareal communication between area 3b and S2 was severely compromised. In summary, recovery of hand grasping behavior could be mediated by collective efforts from SA-like neurons and altered functional circuits that include area 3b and S2 and beyond.

Footnotes

  • This work was supported by Dana Foundation Grant to L.M.C. and NS078680 to J.C.G. R.W. is currently supported by NSFC 31700902 and 31871085; Shanghai Municipal of Science and Technology Project 21ZR1407300, 20JC1419500, and 2018SHZDZX01; ZJ Lab; and Shanghai Center for Brain Science and Brain-Inspired Technology. We thank Dr. Jamie Reed for language editing of the manuscript; and Chaohui Tang for technical support on data acquisition.

  • The authors declare no competing financial interests.

  • Correspondence should be addressed to Li Min Chen at limin.chen{at}vanderbilt.edu

References

  1. ↵
    2. Adjamian P
    (2014) The application of electro- and magneto-encephalography in tinnitus research: methods and interpretations. Front Neurol 5:228. doi:10.3389/fneur.2014.00228 pmid:25431567
    OpenUrlCrossRefPubMed
  2. ↵
    2. Azouz R,
    3. Gray CM
    (2000) Dynamic spike threshold reveals a mechanism for synaptic coincidence detection in cortical neurons in vivo. Proc Natl Acad Sci USA 97:8110–8115. doi:10.1073/pnas.130200797 pmid:10859358
    OpenUrlAbstract/FREE Full Text
  3. ↵
    2. Azouz R,
    3. Gray CM
    (2003) Adaptive coincidence detection and dynamic gain control in visual cortical neurons in vivo. Neuron 37:513–523. doi:10.1016/s0896-6273(02)01186-8 pmid:12575957
    OpenUrlCrossRefPubMed
  4. ↵
    2. Blake DT,
    3. Hsiao SS,
    4. Johnson KO
    (1997) Neural coding mechanisms in tactile pattern recognition: the relative contributions of slowly and rapidly adapting mechanoreceptors to perceived roughness. J Neurosci 17:7480–7489. pmid:9295394
    OpenUrlAbstract/FREE Full Text
  5. ↵
    2. Bors E
    (1979) Extinction and synesthesia in patients with spinal-cord injuries. Paraplegia 17:21–31. doi:10.1038/sc.1979.9 pmid:492748
    OpenUrlCrossRefPubMed
  6. ↵
    2. Buzsáki G,
    3. Anastassiou CA,
    4. Koch C
    (2012) The origin of extracellular fields and currents: EEG, ECoG, LFP and spikes. Nat Rev Neurosci 13:407–420. doi:10.1038/nrn3241 pmid:22595786
    OpenUrlCrossRefPubMed
  7. ↵
    2. Carter AW,
    3. Chen SC,
    4. Lovell NH,
    5. Vickery RM,
    6. Morley JW
    (2014) Convergence across tactile afferent types in primary and secondary somatosensory cortices. Plos One 9:e107617. doi:10.1371/journal.pone.0107617 pmid:25215534
    OpenUrlCrossRefPubMed
  8. ↵
    2. Chen LM,
    3. Friedman RM,
    4. Ramsden BM,
    5. LaMotte RH,
    6. Roe AW
    (2001) Fine-scale organization of SI (area 3b) in the squirrel monkey revealed with intrinsic optical imaging. J Neurophysiol 86:3011–3029. doi:10.1152/jn.2001.86.6.3011 pmid:11731557
    OpenUrlCrossRefPubMed
  9. ↵
    2. Chen LM,
    3. Heider B,
    4. Williams GV,
    5. Healy FL,
    6. Ramsden BM,
    7. Roe AW
    (2002) A chamber and artificial dura method for long-term optical imaging in the monkey. J Neurosci Methods 113:41–49. doi:10.1016/s0165-0270(01)00475-7 pmid:11741720
    OpenUrlCrossRefPubMed
  10. ↵
    2. Chen LM,
    3. Qi HX,
    4. Kaas JH
    (2012) Dynamic reorganization of digit representations in somatosensory cortex of nonhuman primates after spinal cord injury. J Neurosci 32:14649–14663. doi:10.1523/JNEUROSCI.1841-12.2012 pmid:23077051
    OpenUrlAbstract/FREE Full Text
  11. ↵
    2. Fisher KM,
    3. Garner JP,
    4. Darian-Smith C
    (2020) Reorganization of the primate dorsal horn in response to a deafferentation lesion affecting hand function. J Neurosci 40:1625–1639. doi:10.1523/JNEUROSCI.2330-19.2020 pmid:31959698
    OpenUrlAbstract/FREE Full Text
  12. ↵
    2. Florence SL,
    3. Jain N,
    4. Kaas JH
    (1997) Plasticity of somatosensory cortex in primates. Semin Neurosci 9:3–12. doi:10.1006/smns.1997.0101
    OpenUrlCrossRef
  13. ↵
    2. Freund P,
    3. Curt A,
    4. Friston K,
    5. Thompson A
    (2013) Tracking changes following spinal cord injury: insights from neuroimaging. Neuroscientist 19:116–128. doi:10.1177/1073858412449192 pmid:22730072
    OpenUrlCrossRefPubMed
  14. ↵
    2. Gao L,
    3. Kostlan K,
    4. Wang Y,
    5. Wang X
    (2016) Distinct subthreshold mechanisms underlying rate-coding principles in primate auditory cortex. Neuron 91:905–919. doi:10.1016/j.neuron.2016.07.004 pmid:27478016
    OpenUrlCrossRefPubMed
  15. ↵
    2. Handler A,
    3. Ginty DD
    (2021) The mechanosensory neurons of touch and their mechanisms of activation. Nat Rev Neurosci 22:521–537. doi:10.1038/s41583-021-00489-x pmid:34312536
    OpenUrlCrossRefPubMed
  16. ↵
    2. Hasenstaub A,
    3. Shu Y,
    4. Haider B,
    5. Kraushaar U,
    6. Duque A,
    7. McCormick DA
    (2005) Inhibitory postsynaptic potentials carry synchronized frequency information in active cortical networks. Neuron 47:423–435. doi:10.1016/j.neuron.2005.06.016 pmid:16055065
    OpenUrlCrossRefPubMed
  17. ↵
    2. Hwang EJ,
    3. Andersen RA
    (2011) Effects of visual stimulation on LFPs, spikes, and LFP-spike relations in PRR. J Neurophysiol 105:1850–1860. doi:10.1152/jn.00802.2010 pmid:21307325
    OpenUrlCrossRefPubMed
  18. ↵
    2. Jain N,
    3. Catania KC,
    4. Kaas JH
    (1997) Deactivation and reactivation of somatosensory cortex after dorsal spinal cord injury. Nature 386:495–498. doi:10.1038/386495a0 pmid:9087408
    OpenUrlCrossRefPubMed
  19. ↵
    2. Jain N,
    3. Qi HX,
    4. Collins CE,
    5. Kaas JH
    (2008) Large-scale reorganization in the somatosensory cortex and thalamus after sensory loss in macaque monkeys. J Neurosci 28:11042–11060. doi:10.1523/JNEUROSCI.2334-08.2008 pmid:18945912
    OpenUrlAbstract/FREE Full Text
  20. ↵
    2. Jurkiewicz MT,
    3. Mikulis DJ,
    4. McIlroy WE,
    5. Fehlings MG,
    6. Verrier MC
    (2007) Sensorimotor cortical plasticity during recovery following spinal cord injury: a longitudinal fMRI study. Neurorehabil Neural Repair 21:527–538. doi:10.1177/1545968307301872 pmid:17507643
    OpenUrlCrossRefPubMed
  21. ↵
    2. Kaas JH,
    3. Nelson RJ,
    4. Sur M,
    5. Dykes RW,
    6. Merzenich MM
    (1984) The somatotopic organization of the ventroposterior thalamus of the squirrel monkey, Saimiri sciureus. J Comp Neurol 226:111–140. doi:10.1002/cne.902260109 pmid:6736292
    OpenUrlCrossRefPubMed
  22. ↵
    2. Katzner S,
    3. Nauhaus I,
    4. Benucci A,
    5. Bonin V,
    6. Ringach DL,
    7. Carandini M
    (2009) Local origin of field potentials in visual cortex. Neuron 61:35–41. doi:10.1016/j.neuron.2008.11.016 pmid:19146811
    OpenUrlCrossRefPubMed
  23. ↵
    2. Khamechian MB,
    3. Kozyrev V,
    4. Treue S,
    5. Esghaei M,
    6. Daliri MR
    (2019) Routing information flow by separate neural synchrony frequencies allows for 'functionally labeled lines' in higher primate cortex. Proc Natl Acad Sci USA 116:12506–12515. doi:10.1073/pnas.1819827116 pmid:31147468
    OpenUrlAbstract/FREE Full Text
  24. ↵
    2. Krishna V,
    3. Andrews H,
    4. Varma A,
    5. Mintzer J,
    6. Kindy MS,
    7. Guest J
    (2014) Spinal cord injury: how can we improve the classification and quantification of its severity and prognosis? J Neurotrauma 31:215–227. doi:10.1089/neu.2013.2982 pmid:23895105
    OpenUrlCrossRefPubMed
  25. ↵
    2. Lee J,
    3. Lisberger SG
    (2013) Gamma synchrony predicts neuron-neuron correlations and correlations with motor behavior in extrastriate visual area MT. J Neurosci 33:19677–19688. doi:10.1523/JNEUROSCI.3478-13.2013 pmid:24336731
    OpenUrlAbstract/FREE Full Text
  26. ↵
    2. Lee SH,
    3. Dan Y
    (2012) Neuromodulation of brain states. Neuron 76:209–222. doi:10.1016/j.neuron.2012.09.012 pmid:23040816
    OpenUrlCrossRefPubMed
  27. ↵
    2. Liu J,
    3. Newsome WT
    (2006) Local field potential in cortical area MT: stimulus tuning and behavioral correlations. J Neurosci 26:7779–7790. doi:10.1523/JNEUROSCI.5052-05.2006 pmid:16870724
    OpenUrlAbstract/FREE Full Text
  28. ↵
    2. Long KH,
    3. Lieber JD,
    4. Bensmaia SJ
    (2022) Texture is encoded in precise temporal spiking patterns in primate somatosensory cortex. Nat Commun 13:1311. doi:10.1038/s41467-022-28873-w
    OpenUrlCrossRef
  29. ↵
    2. Mathis A,
    3. Mamidanna P,
    4. Cury KM,
    5. Abe T,
    6. Murthy VN,
    7. Mathis MW,
    8. Bethge M
    (2018) DeepLabCut: markerless pose estimation of user-defined body parts with deep learning. Nat Neurosci 21:1281–1289. doi:10.1038/s41593-018-0209-y pmid:30127430
    OpenUrlCrossRefPubMed
  30. ↵
    2. McCormick DA,
    3. McGinley MJ,
    4. Salkoff DB
    (2015) Brain state dependent activity in the cortex and thalamus. Curr Opin Neurobiol 31:133–140. doi:10.1016/j.conb.2014.10.003 pmid:25460069
    OpenUrlCrossRefPubMed
  31. ↵
    2. Niu JW,
    3. Ding L,
    4. Li JJ,
    5. Kim H,
    6. Liu JK,
    7. Li HP,
    8. Moberly A,
    9. Badea TC,
    10. Duncan ID,
    11. Son YJ,
    12. Scherer SS,
    13. Luo WQ
    (2013) Modality-based organization of ascending somatosensory axons in the direct dorsal column pathway. J Neurosci 33:17691–17709. doi:10.1523/JNEUROSCI.3429-13.2013 pmid:24198362
    OpenUrlAbstract/FREE Full Text
  32. ↵
    2. Pei YC,
    3. Denchev PV,
    4. Hsiao SS,
    5. Craig JC,
    6. Bensmaia SJ
    (2009) Convergence of submodality-specific input onto neurons in primary somatosensory cortex. J Neurophysiol 102:1843–1853. doi:10.1152/jn.00235.2009 pmid:19535484
    OpenUrlCrossRefPubMed
  33. ↵
    2. Pomme S
    (2015) The uncertainty of the half-life. Metrologia 52:S51–S65.
    OpenUrl
  34. ↵
    2. Qi HX,
    3. Chen LM,
    4. Kaas JH
    (2011) Reorganization of somatosensory cortical areas 3b and 1 after unilateral section of dorsal columns of the spinal cord in squirrel monkeys. J Neurosci 31:13662–13675. doi:10.1523/JNEUROSCI.2366-11.2011 pmid:21940457
    OpenUrlAbstract/FREE Full Text
  35. ↵
    2. Qi HX,
    3. Gharbawie OA,
    4. Wynne KW,
    5. Kaas JH
    (2013) Impairment and recovery of hand use after unilateral section of the dorsal columns of the spinal cord in squirrel monkeys. Behav Brain Res 252:363–376. doi:10.1016/j.bbr.2013.05.058 pmid:23747607
    OpenUrlCrossRefPubMed
  36. ↵
    2. Qi HX,
    3. Reed JL,
    4. Wang F,
    5. Gross CL,
    6. Liu X,
    7. Chen LM,
    8. Kaas JH
    (2021) Longitudinal fMRI measures of cortical reactivation and hand use with and without training after sensory loss in primates. Neuroimage 236:118026. doi:10.1016/j.neuroimage.2021.118026 pmid:33930537
    OpenUrlCrossRefPubMed
  37. ↵
    2. Reed JL,
    3. Pouget P,
    4. Qi HX,
    5. Zhou Z,
    6. Bernard MR,
    7. Burish MJ,
    8. Haitas J,
    9. Bonds AB,
    10. Kaas JH
    (2008) Widespread spatial integration in primary somatosensory cortex. Proc Natl Acad Sci USA 105:10233–10237. doi:10.1073/pnas.0803800105 pmid:18632579
    OpenUrlAbstract/FREE Full Text
  38. ↵
    2. Roelfsema PR,
    3. Engel AK,
    4. Konig P,
    5. Singer W
    (1997) Visuomotor integration is associated with zero time-lag synchronization among cortical areas. Nature 385:157–161. doi:10.1038/385157a0 pmid:8990118
    OpenUrlCrossRefPubMed
  39. ↵
    2. Saal HP,
    3. Bensmaia SJ
    (2014) Touch is a team effort: interplay of submodalities in cutaneous sensibility. Trends Neurosci 37:689–697. doi:10.1016/j.tins.2014.08.012 pmid:25257208
    OpenUrlCrossRefPubMed
  40. ↵
    2. Schoffelen JM,
    3. Oostenveld R,
    4. Fries P
    (2005) Neuronal coherence as a mechanism of effective corticospinal interaction. Science 308:111–113. doi:10.1126/science.1107027 pmid:15802603
    OpenUrlAbstract/FREE Full Text
  41. ↵
    2. Shu Y,
    3. Hasenstaub A,
    4. Badoual M,
    5. Bal T,
    6. McCormick DA
    (2003) Barrages of synaptic activity control the gain and sensitivity of cortical neurons. J Neurosci 23:10388–10401. pmid:14614098
    OpenUrlAbstract/FREE Full Text
  42. ↵
    2. Smith JE,
    3. Beliveau V,
    4. Schoen A,
    5. Remz J,
    6. Zhan CA,
    7. Cook EP
    (2015) Dynamics of the functional link between area MT LFPs and motion detection. J Neurophysiol 114:80–98. doi:10.1152/jn.00058.2015 pmid:25948867
    OpenUrlCrossRefPubMed
  43. ↵
    2. Suresh AK,
    3. Greenspon CM,
    4. He Q,
    5. Rosenow JM,
    6. Miller LE,
    7. Bensmaia SJ
    (2021) Sensory computations in the cuneate nucleus of macaques. Proc Natl Acad Sci USA 118:e2115772118.
    OpenUrlAbstract/FREE Full Text
  44. ↵
    2. Varela F,
    3. Lachaux JP,
    4. Rodriguez E,
    5. Martinerie J
    (2001) The brainweb: phase synchronization and large-scale integration. Nat Rev Neurosci 2:229–239. doi:10.1038/35067550 pmid:11283746
    OpenUrlCrossRefPubMed
  45. ↵
    2. Wang Z,
    3. Qi HX,
    4. Kaas JH,
    5. Roe AW,
    6. Chen LM
    (2013) Functional signature of recovering cortex: dissociation of local field potentials and spiking activity in somatosensory cortices of spinal cord injured monkeys. Exp Neurol 249:132–143. doi:10.1016/j.expneurol.2013.08.013 pmid:24017995
    OpenUrlCrossRefPubMed
  46. ↵
    2. Weber AI,
    3. Saal HP,
    4. Lieber JD,
    5. Cheng JW,
    6. Manfredi LR,
    7. Dammann JF 3rd.,
    8. Bensmaia SJ
    (2013) Spatial and temporal codes mediate the tactile perception of natural textures. Proc Natl Acad Sci USA 110:17107–17112. doi:10.1073/pnas.1305509110 pmid:24082087
    OpenUrlAbstract/FREE Full Text
  47. ↵
    2. Womelsdorf T,
    3. Fries P
    (2006) Neuronal coherence during selective attentional processing and sensory-motor integration. J Physiol Paris 100:182–193. doi:10.1016/j.jphysparis.2007.01.005 pmid:17317118
    OpenUrlCrossRefPubMed
  48. ↵
    2. Womelsdorf T,
    3. Fries P,
    4. Mitra PP,
    5. Desimone R
    (2006) Gamma-band synchronization in visual cortex predicts speed of change detection. Nature 439:733–736. doi:10.1038/nature04258 pmid:16372022
    OpenUrlCrossRefPubMed
  49. ↵
    2. Wu R,
    3. Su L,
    4. Yang PF,
    5. Min Chen L
    (2016) Altered spatiotemporal dynamics of cortical activation to tactile stimuli in somatosensory area 3b and area 1 of monkeys after spinal cord injury. eNeuro 3:ENEURO.0095-16.2016. doi:10.1523/ENEURO.0095-16.2016
    OpenUrlAbstract/FREE Full Text
  50. ↵
    2. Wu R,
    3. Yang PF,
    4. Chen LM
    (2017) Correlated disruption of resting-state fMRI, LFP, and spike connectivity between area 3b and s2 following spinal cord injury in monkeys. J Neurosci 37:11192–11203. doi:10.1523/JNEUROSCI.2318-17.2017 pmid:29038239
    OpenUrlAbstract/FREE Full Text
  51. ↵
    2. Yang PF,
    3. Qi HX,
    4. Kaas JH,
    5. Chen LM
    (2014) Parallel functional reorganizations of somatosensory areas 3b and 1, and S2 following spinal cord injury in squirrel monkeys. J Neurosci 34:9351–9363. doi:10.1523/JNEUROSCI.0537-14.2014 pmid:25009268
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Email
Print
View Full Page PDF
Citation Tools
Respond to this article
Differential Recovery of Submodality Touch Neurons and Interareal Communication in Sensory Input-Deprived Area 3b and S2 Cortices
Ruiqi Wu, Pai-Feng Yang, Feng Wang, Qing Liu, John C. Gore, Li Min Chen
Journal of Neuroscience 14 December 2022, 42 (50) 9330-9342; DOI: 10.1523/JNEUROSCI.0034-22.2022
Twitter logo Facebook logo Mendeley logo

Jump to section

Keywords

Responses to this article

Respond to this article

Jump to comment:

No eLetters have been published for this article.