(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Posteromedial thalamic nucleus activity significantly contributes to perceptual discrimination [1]

['Jia Qi', 'Unit On Functional Neural Circuits', 'National Institute Of Mental Health', 'National Institutes Of Health', 'Bethesda', 'Maryland', 'United States Of America', 'Changquan Ye', 'Shovan Naskar', 'Ana R. Inácio']

Date: 2022-12 

(A) Schematic of texture discrimination task in freely moving mice. Upper, top view cartoon demonstrates the sequence of the self-initiated whisker-dependent texture discrimination 2AFC task. Lower, top view image of a mouse with the optical fibers approaching sensory zone. The color dots are post hoc labeled by DeepLabCut tracking software to indicate the mouse’s nose, ears, back, and tail. Left, task sequence. (B) Schematic of texture presentation. Left, behavioral chamber design with nose poking sensors, retractable window, two rotatable cubes with four different textures on four surfaces of each cube, and water reward ports. Right, four different grits of sandpapers (P120, P180, P280, and P400) are mounted on each cube. Two cubes present two different textures in each trial. (C) Learning curve for the mice trained on texture discrimination task (texture P120 vs. P400). Gray line represents a learning curve for individual mouse (n = 28) over multiple sessions. The thick black line represents average performance across mice, and the thin black line indicates SEM. Blue dotted line indicates 70% of correct trials in task performance. (D) Dependence of task performance on texture difference. Texture difference index (TDI) is calculated as the difference of two texture grits divided by 100. Task performance increases as TDI increases. (E) Behavioral performance decreases to chance level after bilateral whisker trimming (P < 0.0001, n = 28, paired t test). Gray bars indicate average across 28 mice, and black lines represent an individual mouse. The data in Fig 1C–1E, can be found in S1 Data .

To investigate the role of POm in sensory perception during natural behavioral conditions, we designed a whisker-dependent texture discrimination task in the form of a self-initiated, 2AFC task in freely moving mice ( Fig 1A and 1B ). Animals initiated each trial by poking their nose into the center port of a task chamber. Following trial initiation, two different textures were simultaneously presented for 1 s. Animals were trained to sense two different textures with their whiskers and to choose one of the two side ports for a reward that was associated with a target texture. There was no punishment for incorrect choices or time-out before the animals initiated a subsequent trial. The animal’s behavior was monitored and video recorded during the task. Sandpapers with four different grits (P120, P180, P280, and P400) were used for the texture discrimination task ( Fig 1B ). The four sandpapers with different grits were mounted on a cube, and two cubes presented two different textures for each trial. Initially, animals were trained to discriminate two different textures, P120 and P400. Once the animals reached performance accuracy of 70% on three consecutive sessions, the animals were considered to have learned the task ( Fig 1C ). Then, the target texture (P120 or P400) was presented with one of the other three different textures in a pseudorandomized order ( Fig 1B ). The target texture was counterbalanced across mice. To describe the different levels of the task difficulty, we employed a texture difference index (TDI; TDI = (average particle size of texture1 − average particle size of texture 2) / 100). Task performance increased as the TDI increased ( Fig 1D ). To avoid the influence of visual cues, the behavioral task was conducted in a dark environment. To further confirm that the animals use their whiskers to perform the task, we bilaterally trimmed all the whiskers at the end of all experiments. After trimming, the animals performed the task at chance level, indicating that the animals rely on their whiskers to perform the task ( Fig 1E ; before: 82.15 ± 1.10% versus after: 52.97 ± 0.74%; t = 22.39, df = 27, n = 28, P < 0.0001). These results demonstrate that our self-initiated texture discrimination task is a simple yet robust behavior task to study whisker-dependent sensory perception in freely moving animals.

(A) Schematic of experimental design and a representative image showing the expression of hM4D(Gi)-mCherry in POm. Scale bar = 500 μm. (B) Chemogenetic suppression of POm significantly impairs the task performance of texture discrimination task. CNO injection, but not saline injection, significantly attenuates the behavior performance of the hM4D(Gi)-mCherry group mice in the whisker-dependent 2AFC task (P < 0.0001, n = 16, paired t test). In the AAV-hSyn-mCherry control group mice, CNO injection shows no effect on task performance (P = 0.38, n = 6, paired t test). ( C ) Schematic of experimental design and a representative image showing the expression of hM4D(Gi)-mCherry in VPm. ( D ) Chemogenetic suppression of VPm results in chance level performance. (P < 0.0001, n = 6, paired t test). Scale bar = 500 μm. The data in Fig 2B and 2E can be found in S1 Data . CNO, clozapine-N-oxide; POm, posteromedial; VPm, ventral posteromedial nucleus; 2AFC, two-alternative forced choice.

POm neurons show slower and weaker responses to whisker stimulation, while ventral posteromedial nucleus (VPm) neurons, a first-order thalamic nucleus of the somatosensory system, strongly respond to passive whisker stimulation with high spatiotemporal fidelity [ 26 , 31 , 32 ]. Considering that POm receives converging inputs from diverse brain areas, weak responsiveness of POm neurons raises questions on the role of POm in sensory perception, and behavioral conditions and stimuli that engage POm neurons. However, most studies were conducted using passive sensory stimulation under anesthetized or head-fixed conditions, thus making it difficult to understand the function of POm during active sensing in awake animals. To address this question, we suppressed POm activity with a chemogenetic method using designer receptors exclusively activated by designer drugs (DREADDs) while the freely moving animals were performing the self-initiated texture discrimination task ( Figs 2 and S1 ). We first tested the efficacy of clozapine-N-oxide (CNO) on hM4D(Gi)-expressing POm neurons. We confirmed that CNO application significantly lowered the resting membrane potential and increased the rheobase of hM4D(Gi)-expressing POm neurons ( S2 Fig , resting membrane potential, control −67.93 ± 1.04 Vm versus CNO −70.95 ± 1.34 Vm, P = 0.0059; rheobase, control 33.00 ± 7.53 pA versus CNO 69.36 ± 13.28 pA, P = 0.0020, n = 11 cells from 4 mice). We bilaterally injected inhibitory DREADDs (AAV-hSyn-hM4D(Gi)-mCherry) [ 33 ] into POm three weeks prior to the behavioral task. We then interleaved the behavioral task sessions with CNO (3 mg/kg, intraperitoneally (IP)) or saline injection ( Fig 2A ). The direct suppression of POm neurons with chemogenetic approach significantly impaired the animal’s performance in the texture discrimination task ( Fig 2B ; saline 83.09 ± 1.90% versus CNO 70.55 ± 2.38%, t = 16.15, df = 15, P <0.0001, n = 16, paired t test). To test the off-target effect of CNO, we also injected a control virus (AAV-hSyn-mCherry) into POm in a different set of animals and administered CNO or saline. Task performance between saline and CNO-injected groups did not differ ( Fig 2B ; saline 80.14 ± 1.04% versus CNO 79.03 ± 1.83%, t = 0.96, df = 5, P = 0.38, n = 6, paired t test). While the suppression of POm activity significantly lowered the rate of correct response, the animals’ performance was still higher than chance levels. This is because VPm was still intact. To test the role of VPm in sensory perception tasks, we bilaterally injected inhibitory DREADDs into VPm ( Fig 2C ). As expected, chemogenetic suppression of VPm resulted in chance level performance ( Fig 2D , P < 0.0001, n = 6, paired t test). However, the significant role of POm in sensory discrimination is unexpected given the slow and weak responsiveness of POm neurons to passive sensory stimulation and the low spatial resolution of POm receptive fields [ 26 , 31 , 32 ]. To ensure that the expression of hM4D(Gi) receptors was restricted within POm and did not spread to VPm, we validated the expression of hM4D(Gi)-mCherry in POm and VPm from each mouse ( S3 Fig ). While our injection was mostly restricted within POm and reliably avoided VPm, we consistently detected the expression of hM4D(Gi)-mCherry in the lateral posterior (LP) thalamic nucleus that is situated just above POm ( Fig 2A ). LP, the higher-order thalamic nucleus in visual system of rodents, is highly sensitive to diverse behavioral contexts [ 5 , 6 , 34 – 36 ]. To address whether the expression of hM4D(Gi) receptors in LP affect the animals’ performance during the texture discrimination task, we bilaterally expressed hM4D(Gi) receptors only in LP but not in POm ( S4 Fig ). Suppression of LP activity with CNO injection did not affect the task performance, demonstrating that the impairment in the animals’ performance in the texture discrimination task is due to the suppression of POm activity and not because of the alteration of LP activity ( S4 Fig ; saline 81.40 ± 1.37% versus CNO 80.30 ± 1.26%; t = 1.32, df = 5, P = 0.24; n = 6, paired t test). In summary, our results demonstrate that POm, a higher-order thalamic nucleus in the somatosensory system, plays a significant role in texture discrimination during active sensing.

Corticothalamic inputs from S1 are necessary for the role of POm in perceptual sensory discrimination

We next asked which input sources are critical for the contribution of POm in perceptual sensory discrimination. POm receives not only afferents originating from the brainstem, mostly from spinal trigeminal interpolaris nucleus (SP5i), through the paralemniscal pathway [13,14] but also inputs from multiple cortical areas, including motor and somatosensory cortices via the descending pathway [13,17–19,37]. To identify the main excitatory input source that enables POm to play a significant role in sensory perception, we manipulated the afferents to POm with an optogenetic approach. We asked whether the suppression of the axon terminals from the major afferents, SP5i, M1/M2, and S1 to POm affect the sensory perception of the animals during the texture discrimination task (Fig 3).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Corticothalamic inputs from S1 are necessary for the role of POm in perceptual sensory discrimination. (A, B, C, E) Experimental strategy with chemogenetic and optogenetic manipulation of POm and inputs to POm. Left, schematic of experimental design. Right, representative images showing the expression of hM4D(Gi)-mCherry in POm (A, C, and E) and eNpHR3.0-eYFP in Sp5i (A), M1/M2 (C), and S1 (E). In three different groups of Rbp4-cre mice, Cre-dependent eNpHR-eYFP was injected into L5 of S1 or M1/M2, while AAV-hSyn-eNpHR- eYFP was delivered into SP5i. In all three groups, AAV-hSyn-hM4D(Gi)-mCherry was injected into POm. All of the viral injections were bilateral, and the optical probes were bilaterally implanted above POm. (B) Role of Sp5i inputs to POm in sensory discrimination. Left, optogenetic suppression of SP5i axon terminals in POm does not affect task performance (P = 0.22, n = 5, paired t test), while chemogenetic inhibition of POm (right) or the combined inhibition (middle) significantly reduce task performance (P = 0.0002, n = 5, paired t test). (D) Role of M1/M2 inputs to POm in sensory discrimination. Left, optogenetic suppression of M1/M2 axon terminals in POm does not affect task performance (P = 0.84, n = 4, paired t test), while chemogenetic suppression of POm (right) or the combined suppression (middle) significantly reduce task performance (P = 0.0002, n = 4, paired t test). (F) Role of S1 inputs to POm in sensory discrimination. Upper, schematic of the optogenetic (left), chemogenetic (right), or the combined (middle) manipulation strategies. Left, optogenetic suppression of S1 axon terminals in POm significantly attenuates task performance (P = 0.0002, n = 7, paired t test). Decrease in task performance during chemogenetic suppression of POm (right, P < 0.0001, n = 7, paired t test) or combined both optogenetic and chemogenetic suppression (middle, P < 0.0001, n = 7, paired t test) is comparable to the level of the attenuation during optogenetic suppression of S1 axon terminals innervating POm. The gray bar indicates the average across mice, and the black line represents an individual mouse. (G) AAV5-EF1a-DIO-eYFP were bilaterally injected into S1 and AAV5-CAG-mCherry into POm of control group (Rbp4-Cre). Neither optogenetic nor combined suppression had effects on task performance in the control group (left panel, P = 0.6629; right panel, P = 0.0695; n = 6, paired t test). (H) Suppression of corticothalamic inputs from S1 to POm significantly lowers sensory discrimination. Normalized correct choice of three groups in three different manipulations. While chemogenetic suppression of POm significantly reduces correct choice in all three groups, only optogenetic suppression of S1 inputs to POm affects task performance similar to POm suppression, but not the M1/M2 and Sp5i inputs to POm groups. Scale bar = 500 μm in (A, C, and E). The data in Fig 3B, 3D, 3F, 3G, and 3H can be found in S1 Data. https://doi.org/10.1371/journal.pbio.3001896.g003

To target corticothalamic inputs from S1 and M1/M2 to POm, we employed an Rbp4-cre mouse line in which Cre is expressed in layer 5 (L5) pyramidal neurons [38], the main cortical layer that projects to POm [37]. We bilaterally injected AAV vectors expressing halorhodopsin (AAV5-EF1a-DIO-eNpHR3.0-eYFP) [39] into motor cortex (M1/M2), and primary somatosensory cortex (S1) in separate animals (Fig 3C and 3E). To minimize mouse strain-dependent effects on mouse behavior, we used the Rbp4-cre mouse line in targeting SP5i neurons despite constitutive AAV vectors (AAV5-hSyn-eNpHR3.0-eYFP) were injected to SP5i (Fig 3A). We also bilaterally expressed hM4D(Gi) receptors in POm of these three groups of mice. This strategy allowed us to independently control the activity of POm neurons chemogenetically and the excitatory inputs to POm optogenetically. Optical fibers (200 μm) were bilaterally implanted above POm to deliver 590 nm light (8 to 10 mW) for optogenetic suppression of the axon terminals from the main afferents (Fig 3A, 3C, and 3E). The light application was triggered by trial initiation and terminated once an animal made a choice by poking its nose into the water port. To confirm the optogenetic suppression of axon terminals innervating POm, we performed in vivo extracellular recording from POm using an optrode (optic fiber 100 μm) (S5 Fig). We recorded spontaneous activity of POm neurons, while applying optogenetic stimulation to POm innervating M1/M2 terminals in awake, head-fixed mice. Optogenetic application significantly suppressed spontaneous spiking activity (21 out of 109 POm units, baseline 7.49 ± 1.25 Hz versus 590 nm ON 4.99 ± 0.95 Hz, P < 0.001, paired t test).

Bilateral optogenetic suppression of SP5i axon terminals in POm showed no effect on the rate of correct response (Fig 3B; LED off 80.36 ± 0.59% versus LED on 79.50 ± 0.51%, t = 1.456, df = 4, P = 0.22, n = 5, paired t test), while chemogenetic suppression of POm neurons with CNO injection in the same animals did significantly lower the task performance (Fig 3B; saline 82.25 ± 0.60% versus CNO 68.82 ± 0.42%, t = 44.04, df = 4, P < 0.0001, n = 5, paired t test). Expression of halorhodopsin was localized in the SP5i and was not detected in the principal trigeminal nucleus (Pr5), which is the main brainstem nucleus for lemniscal pathway. Consistent with previous studies [16,40,41], we detected the innervation of SP5i both in POm and in VPm (Fig 3A).

Similar to SP5i inputs to POm, suppression of the axon terminals from M1/M2 did not change the animals’ performance compared to the light-off conditions (Fig 3D; LED off 87.96 ± 1.54% versus LED on 87.67 ± 1.41%, t = 0.2139, df = 3, P = 0.84, n = 4, paired t test). However, CNO-induced suppression of POm activity in the same animals significantly lowered task performance (Fig 3D; saline 84.45 ± 1.82% versus CNO 74.25 ± 1.51%, t = 4.387, df = 3, P = 0.022, n = 4, paired t test). This result demonstrates that inputs from SP5i and M1/M2 do not transmit perceptual information to POm. We wondered whether cre-dependent expression of NpHR within a Rbp4-cre mouse line may limit the expression of NpHR only within a subpopulation of M1/M2 L5/6 cells projecting to POm and, thus, result in the lack of optogenetic suppression effect in task performance. We first examined the laminar distribution of POm-projecting neurons by injecting AAVretro-CAG-Cre to POm of an Ai14 mouse (S6 Fig). While we detected POm-projecting neurons from both L5 and L6, we found more abundant POm-projecting cells in L5 (S6 Fig). To further test the role of M1/M2 inputs to POm in texture discrimination, we constitutively expressed NpHR from neurons in M1/M2 of wild-type mice. This allows more broad and unbiased expression of NpHR in M1/M2 cells. Optogenetic suppression of these inputs to POm resulted in no detectable impact on task performance, suggesting that M1/M2 inputs to POm do not play an important role in sensory discrimination (S6 Fig; P = 0.43, n = 4, paired t test).

In contrast to SP5i and M1/M2 inputs, suppression of cortical inputs from S1 to POm (Fig 3E) significantly reduced task performance (Fig 3F; LED off 84.89 ± 2.46% versus LED on 72.63 ± 2.05%, t = 8.272, df = 6, P = 0.0002, n = 7, paired t test). The reduction in the rate of correct responses during optogenetic perturbation of S1 inputs to POm was similar to the decrease when the activity of POm was suppressed with CNO in the same animals (Fig 3F; saline 82.93 ± 2.30% versus CNO 69.65 ± 2.92%, t = 12.3, df = 6, P < 0.0001, n = 7, paired t test). To further test whether S1 inputs are the main source in conveying the perceptual information during the texture discrimination task, we suppressed POm activity with CNO injection, in addition to the optogenetic perturbation of S1 inputs. The combined perturbation did not further reduce performance (Fig 3F; LED off + saline 83.93 ± 2.47% versus LED on + CNO 69.47 ± 2.51%, t = 14.61, df = 6, P < 0.0001, n = 7, paired t test). Moreover, no changes in behavioral performance were found in the control group (Fig 3G; LED off 80.21 ± 1.57% versus LED on 79.67 ± 1.45%, t = 0.4629, df = 5, P = 0.66, n = 6, paired t test; LED off + saline 79.65 ± 1.73% versus LED on + CNO 78.52 ± 1.93%, t = 2.303, df = 5, P = 0.0695, n = 6, paired t test). The normalized behavioral performance (Fig 3H) shows that the role of POm in sensory discrimination is mainly inherited from primary sensory cortical projections but not from motor cortical projections or brainstem ascending pathway. Together, our finding suggests that the cortical inputs from S1 are mostly responsible for the role of POm in perceptual sensory discrimination.

To examine whether CNO and optogenetic suppression alters the animals’ motivation and movement, and consequent impairment in task performance, we analyzed behavioral parameters of the sensory discrimination task. We analyzed total trial numbers in each session, trial duration, intertrial interval, movement velocity, and duration of active sensing against the presented textures in each trial using DeepLabCut [42]. These parameters under optogenetic and chemogenetic manipulation conditions were comparable to those under control conditions (S7 Fig). The results suggest that the impaired performance in the discrimination task was not due to movement deficit or the animals’ motivation.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001896

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/