Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2023 Jan 13.
Published in final edited form as: J Cogn Neurosci. 2022 Jun 2;34(7):1259–1273. doi: 10.1162/jocn_a_01855

Few pairings of electrical micro-stimulation of the ventral tegmental area and a visual stimulus enhance visual cortical responses

Jerome Herpers 1,2, Wim Vanduffel 1,2,3,4, Rufin Vogels 1,2,5
PMCID: PMC7614035  EMSID: EMS158944  PMID: 35468206

Abstract

Previous studies demonstrated that pairing a visual stimulus and electrical microstimulation of the ventral tegmental area (VTA-EM) for multiple days is sufficient to induce visual cortical plasticity and changes perception. However, a brief epoch of VTA-EM stimulus pairing within a single day has been shown to result in a behavioral preference for the paired stimulus. Here, we investigated whether a brief single-day session of VTA-EM-stimulus pairings is sufficient to induce changes in visual cortical responses. We examined macaque posterior inferior temporal cortex (PIT) since previous studies demonstrated response changes after VTA-EM stimulus pairing in that area. Multi-unit recordings in PIT were interleaved with VTA-EM-stimulus pairing epochs. During the short VTA-EM-stimulus pairing epochs (60 pairings), one image (fractal) was paired with VTA-EM (STIM) while another, unpaired fractal was presented as control (CON). Two other fractals (dummies) were presented only during the recordings. The difference in response between the STIM and CON fractals already increased after the first VTA-EM-stimulus pairing epoch, reflecting a relative increase of the response to the STIM fractal. However, the response to the STIM fractal did not increase further with more VTA-EM-stimulus pairing epochs. The relative increase in firing rate for the paired fractal was present early in the response, in line with a local/bottom-up origin. These effects were absent when comparing the responses to the dummies pre-and post-VTA-EM. This study shows that pairing a visual image and VTA-EM in a brief single-day session is sufficient to increase the response for the paired image in macaque PIT.

Introduction

To survive in a continuously changing environment, organisms have to adapt their behavior accordingly. To tend to these changes, the adult brain is increasing or diminishing synaptic strengths amongst neurons. This process is known as adult cortical plasticity (Buonomano & Merzenich, 1998; Gilbert & Li, 2012). Plastic changes can be produced gradually and can persist over a long time, sometimes even lifelong (Draganski et al., 2004; Ungerleider et al., 2002).

One of the brain areas supporting plastic changes is the ventral tegmental area (VTA). A vast amount of research in both rodents and primates has provided evidence that dopaminergic neurons of the VTA (Margolis et al., 2006) encode reward prediction errors, which play a causal role in different types of learning and associated cortical plasticity (for recent reviews, see: Lerner et al., 2021; Schultz, 2019; Schultz et al., 2017). Increasing the ability to detect rewarding events, objects, or features by learning increases the chances to receive more rewards in the future. Hence, dopamine can strengthen active synaptic connections when rewards are greater than expected (for reviews e.g.: Berke, 2018; Glimcher, 2011; Jay, 2003; Pignatelli & Bonci, 2015; Speranza et al., 2021). Note that dopaminergic VTA neurons might be involved in other functions than encoding reward prediction errors because some of these neurons may also respond to aversive stimuli (de Jong et al., 2019) and show movement-related responses (Engelhard et al., 2019; Hughes et al., 2020; Lerner et al., 2021; Schultz, 2019). In addition, dopamine released by the ventral midbrain has also been implicated in reward prediction and value signaling which is important for motivational and effort-related behavior (Berridge, 2007; Bromberg-Martin et al., 2010; Hamid et al., 2015; Salamone & Correa, 2012; Vancraeyenest et al., 2020).

Electrical micro-stimulation of the ventral tegmental area (VTA-EM) can have long-lasting effects on cortical responses to stimuli that were paired with VTA-EM (Arsenault & Vanduffel, 2019; Bao et al., 2001; Herpers et al., 2021; Murris et al., 2021). Monkey studies showed that when a task-irrelevant visual feature is paired with VTA-EM, while attention is drawn away from this stimulus, the representation of that feature increases in posterior inferior temporal cortex (PIT), at both fMRI (Arsenault & Vanduffel, 2019; Murris et al., 2021) and spiking activity level (Herpers et al., 2021). This suggests that VTA-EM paired with a visual stimulus is sufficient to induce cortical plasticity in visual areas. Moreover, pharmacological blockage of dopamine receptors during VTA-EM-stimulus pairing suggests that dopaminergic activity is necessary for such plastic changes, at least in rodent auditory cortex (Bao et al., 2001).

These studies employed long periods of VTA-EM-stimulus pairing, ranging from multiple days to weeks. However, it is unknown whether such long periods of VTA-EM pairing are required to obtain changes in responses in sensory cortical areas. We do know that relatively few VTA-EM-stimulus pairing events suffice to bias behavioral preferences of a monkey towards the paired stimulus within a single experimental session (Arsenault et al., 2014; Herpers et al., 2021; this paper), but it is not known whether such a relatively small number of VTA-EM-stimulus pairings also changes sensory responses in cortical areas. In fact, the time course of the effects of VTA-EM-stimulus pairing in primate visual cortex is still unknown.

To examine this, we paired visual stimuli with VTA-EM during passive fixation in brief epochs of ~60 stimulus-VTA-EM pairings, interleaved with multiunit activity (MUA) recordings in macaque PIT. We recorded from this extrastriate visual area, since previous studies (Arsenault & Vanduffel, 2019; Herpers et al., 2021) showed changes in the responses of PIT neurons, yet after extensive VTA-EM stimulus pairings. We employed fractals as visual stimuli since these are effective stimuli to drive PIT neurons. In each daily session, two new fractals were presented in the VTA-EM stimulus pairing epoch, whereby only one of the fractals was paired with VTA-EM. To assess response changes as a result of VTA-EM pairing, we compared MUA in PIT to the same stimuli pre-and post-VTA-EM-stimulus pairing epochs. In addition, 2 other fractals were presented during recording, but not during the VTA-EM epochs, and used for comparison. We found that even after a single epoch of 60 VTA-EM-stimulus pairings, the responses of PIT neurons showed a relatively increased response for the paired compared to the unpaired fractal stimulus. No such effect was observed for the dummy control stimuli. The relatively increased response to the paired stimulus was already evident in the initial phase of the visual response, supporting a bottom-up or local origin. These data suggest that relatively few pairings of a visual image with VTA-EM within a single day are already sufficient to increase the response for the paired image in area macaque PIT.

Methods

Experimental model and subjects

Two male rhesus macaques (Macaca Mulatta), that also participated in the study of (Herpers et al., 2021), were used in this study (M1: ~9kg, 7 years old; M2: ~10kg, 7 years old). Both subjects were trained in fixation tasks for electrical micro-stimulation of the VTA (VTA-EM) and electrophysiological recordings. A polymer headpost was fixed to their skull for a stable head position during VTA-EM and recordings. We received ethical approval for this study from the Animal Ethics Committee of the KU Leuven Medical School (P113-2014).

Electrical micro-stimulation of the VTA (VTA-EM)

Microelectrodes were chronically implanted in the monkey’s midbrain posterior to the substantia nigra, where the VTA is located (Figure 1A). We followed procedures of (Arsenault et al., 2014), using peri-operative MRI guidance for electrode implantation. In preparation, we implanted MRI compatible chambers (Crist Instruments) on the monkeys’ skulls, projecting towards the VTA. Subsequently, we inserted microwire brush electrode arrays (Bondar et al., 2009), consisting of twenty 25–50-micron thin wires that spread out in the targeted area. Electrodes were placed in the right hemisphere (RH) of M1 and the left hemisphere (LH) of M2. Two constant current stimulus isolators (A365, World Precision Instruments) triggered by custom software were used to send EM events through two electrode wires, chosen based on the lowest impedance. EM events consisted of biphasic, square-wave pulses with a repetition rate of 200Hz for M1 and 50Hz for M2 and lasted for 200ms. Each pulse comprised of a 0.2ms positive and 0.2ms negative voltage with a 0.025ms interval. We used a lower stimulation frequency in M2 since VTA-EM evoked a pupil constriction that interfered with the eye-position signal at higher frequencies. VTA-EM evoked pupil dilation in M1 at high EM currents. For determining the current intensity during the VTA-EM-stimulus pairing sessions, we selected a current intensity that was just below the lowest intensity that would elicit pupil dilation in M1 in each daily session, ranging from 300μA-600μA. For M2, we used a current intensity of 1000μA, which still elicited pupil constriction but without affecting the measured eye gaze. The EM-induced constriction in pupil size in M2 might be a result of unintentional stimulation of the oculomotor nerve, which runs anteriorly from the oculomotor nuclei in the center of the midbrain.

Figure 1. VTA-EM-stimulus pairing and the effect on free-viewing behavior.

Figure 1

Open in a new tab

(A) Location VTA microwire brush electrodes. T1 weighted horizontal MRI section. The electrode tips, producing a black dot on the MRI image, are indicated by an arrow and are located in the right hemisphere of M1 and the left hemisphere of M2 in the midbrain posterior to the substantia nigra. (B) Paradigm. A free-viewing task with two fractals (size: 4°) was performed before and after pairing one of the fractals with VTA-EM to investigate whether stimulating the VTA can induce a behavioral preference for the VTA-EM paired fractal. During the free-viewing task, the monkeys could look anywhere on the screen to receive a juice reward every 1.5s, while two fractals were presented left and right, interchanging positions randomly over trials. During the VTA-EM epochs, the fractals from the free-viewing task were presented one-by-one at an eccentricity of 3°. One fractal was paired with VTA-EM while the other fractal was presented without pairing. The monkeys received a juice reward for fixating every 1.5s, regardless of stimulus presentation or VTA-EM timing. Images are not presented at true scale. (C) Effect of VTA-EM on free-viewing behavior. The preference index (PI; [viewing time for the VTA-EM paired fractal – viewing time for the non-paired fractal] / [total viewing time]) was calculated for each pre-and post-VTA-EM free-viewing session, the number of free-viewing epoch pairs is indicated by n. The difference in PI was calculated by subtracting the pre-VTA-EM pairing PI from the post-VTA-EM pairing PI. Median differences in PI post – pre are indicated (▾) on the distributions of PI post – pre for each monkey. Both monkeys consistently developed a preference for the fractal image paired with VTA-EM post-stimulation compared to pre-stimulation (one-sided Wilcoxon signed-rank test, M1: p = 0.015, M2: p = 0.001).

Behavioral paradigm: Free-viewing

In a previous study, we showed that pairing VTA-EM with a visual stimulus presented in the center of the visual field can produce a behavioral preference for the VTA-EM paired image (Herpers et al., 2021). Here, we tested whether pairing VTA-EM with a stimulus presented in the peripheral visual field at 3° eccentricity showed similar effects on behavioral preference. We employed a similar procedure for the free-viewing and VTA-EM pairing tasks as in (Herpers et al., 2021).

During the free-viewing task (Figure 1B), the monkeys received juice rewards every 1.5s whenever the gaze was within the limits of the screen (40° x 30°). A trial consisted of the presentation of two colorful, geometrical fractals (size 4° x 4°; Sakai & Miyashita, 1991), that were presented 3°to the left and the right of the center of the screen on a grey background. In a trial, the fractals were presented for a total of 2s. The periods that the monkey’s gaze did not stay within the limits of the screen were not counted for the 2s presentation time. The inter-trial interval (ITI) was 1s. The left and right positions of the fractal interchanged randomly across trials. Free-viewing sessions were ended after 10 minutes. The eye position of the monkeys was tracked (1000 Hz, EyeLink, SR-Research) to measure the amount of time the monkeys looked at each fractal. The free-viewing task was performed before and after a VTA-EM pairing session, where one of the two fractals was paired with VTA-EM while the other fractal was presented without VTA stimulation.

VTA-EM-stimulus pairing

Following a free-viewing session with a unique fractal pair, a stimulus-VTA-EM association was performed during a 20-minute long fixation task (Figure 1B). To start a trial, the monkeys had to fixate within a window of 2° a red fixation square (0.14°), which was presented in the center of the screen on a grey background. After 3500-5000ms of fixation, one of the fractals from the previous free-viewing task was presented for 500ms at 3° eccentricity in the lower left visual field at 225° polar angle for M1 and the lower right visual field for M2 at 315° polar angle. ITI was 500ms. The order of presentation of the two fractals was chosen pseudo-randomly. One of the two fractals was randomly chosen for VTA-EM pairing (~60 stimulation events). The other fractal used in the free-viewing task was presented during the VTA-EM pairing task without subsequent VTA-EM. For the VTA-EM paired fractal, a 200ms long VTA-EM train started at stimulus offset. Trial timing was designed with the intent to avoid habituation of the VTA-EM response. The interval between successive EM events was on average (±SEM) 17.4s (±0.26s).

Main procedures during recording sessions

To investigate whether pairing a fractal image with VTA-EM could induce changes in the response properties of PIT neurons, MUA recordings during fixation and the presentation of fractals were interleaved with the above described VTA-EM-stimulus pairing task. During each of 15 experimental days, 6 recording epochs and 5 interleaved VTA-EM stimulus pairing epochs were employed in M1. For M2, we had 20 experimental days (1 day was excluded because of technical difficulties), with 7 recording epochs and 6 interleaved VTA-EM epochs per day. Each day a new set of 4 fractals was selected based on the responses of one of the 16 probe channels in a search test in which we recorded MUA to 32 fractals. Each day, a new set of 32 fractals were generated using the algorithm of Sakai & Miyashita (1991). 4 fractals images with similar responses were online selected for the subsequent experiments. All 4 fractals were presented during the fixation task while recording MUA, after which 2 fractals were presented during the VTA-EM epoch. During this task, one fractal image was paired with VTA-EM (STIM) while the other fractal image was presented without VTA-EM as control (CON). The two fractals that were not used during VTA-EM sessions but that were shown during the recording epochs are considered dummy fractals (DUM1 and DUM2) and served as a comparison.

Search test with 32 fractals

The neural response to 32 fractals was recorded with a 16 channel V-Probe (Plexon) in PIT (Figure 2A) and based on the peristimulus time histograms (PSTH) of one recording channel, we selected online 4 fractals that produced similar responses. The monkeys had to fixate a 0.14° red square on a grey background within a 2° x 2° window for juice reward given every 1.5 s of fixation. After fixating for 500ms, one of 32 fractals was presented for 200ms at 3° eccentricity in the lower LVF or RVF, depending on the monkey and contralateral to the recorded hemisphere, during fixation. The 32 fractals were presented with an ITI of 500ms in random order in blocks of 32 trials. No VTA-EM was performed during electrophysiological recordings. To complete a trial and obtain a juice reward the monkey had to fixate for a 500ms pre-stimulus period, the 200ms presentation period of the fractal image, and a 300ms post-stimulus interval.

Figure 2.

Figure 2

Open in a new tab

(A) Estimated recording position in PIT. T1 weighted coronal MRI section. The estimated recording location of one penetration is marked in red and indicated by arrows (B) MUA recordings in PIT interleaved with VTA-EM-stimulus pairing paradigm. To examine the effect of pairing a visual stimulus with VTA-EM on the short term, recordings in PIT were interleaved with VTA-EM pairing epochs. During fixation, 4 fractals (size: 4°) were presented at 3° eccentricity. During the VTA-EM epoch, 2 of the 4 fractals that were used in the recording epochs were presented. One fractal was paired with VTA-EM (STIM; indicated by a red outline), one fractal was presented but not paired with VTA-EM as control (CON; indicated by a blue outline). The two other fractals used in the recording epochs were not presented during the VTA-EM epoch and are considered the dummy fractals (DUM1 and DUM2; indicated by grey outlines). During both tasks, the monkeys received a juice reward every 1.5 seconds for fixating the center of the screen, regardless of stimulus and VTA-EM timing. Images are not presented at true scale.

Recordings with 4 fractals

To examine changes in neural responses to fractals as a result of VTA-EM, the monkeys performed the same fixation task as described above, but with the 4 selected fractals (Figure 2B). After the first recording session, the two fractals eliciting the most similar neural response were selected for the VTA-EM epochs and randomly assigned to the STIM and CON conditions. Within an experimental session, the fixation task was repeated 6 (M1) and 7 (M2) times, interleaved with VTA-EM epochs in between every two recording sessions. The number of unaborted trials per fractal image was on average 39 for M1 and 55 for M2.

VTA-EM-stimulus pairing epoch interleaved with recordings

The VTA-EM pairing epoch was performed as described above, with the 2 out of 4 fractals that were presented during the recording epochs (Figure 2B). One fractal was paired with VTA-EM (STIM) and the other fractal was presented, but not paired with VTA-EM as control (CON). Within a daily session, the VTA-EM pairing task was repeated 5 (M1) and 6 (M2) times and alternated with a recording epoch.

Receptive field mapping

3° checkerboards of 16 equally sized squares were presented for 200ms on a grey background, with a luminance polarity inversion of the checkerboard at 100ms. The location of the checkerboard varied pseudo-randomly in blocks of 84 trials using 42 locations. The locations were based on an invisible grid with a spacing of 3°, which consisted of 2 columns ipsilateral and 4 columns contralateral to the recorded hemisphere, 3 rows in the upper visual field, and 4 rows in the lower visual field. Trials were separated by an ITI of 100ms. The monkeys received a juice reward every 1.5s of fixation, regardless of stimulus timing. The number of presentations per location was on average 17 for both monkeys.

MUA recording

MRI compatible recording chambers (Crist Instruments) were implanted over the posterior temporal visual cortex. Recording positions (Figure 2A) were based on receptive field mapping in this region in the same monkeys during a previous study (Herpers et al., 2021). Standard electrophysiological recording procedures were followed. 16 channel probes (V-Probe, Plexon) were used for MUA recordings, by lowering them through a guide tube that was fixed into a Crist grid. We waited with recording for at least 60 min after insertion of the guide tube and multi-channel probe into the cortex to ensure stable MUA recordings. Offline sorting of spikes was performed with Offline Sorter (Plexon). All spikes recorded by one channel were combined into one unit. We recorded in the hemisphere ipsilateral to the VTA electrode used for EM, which was the right hemisphere (RH) for M1 and the left hemisphere (LH) for M2. In both monkeys, two neighboring grid holes were used for recordings and the depth of the probe was constant during the experimental session. In M1, we recorded ventral to the lower bank of the superior temporal sulcus (STS) and dorsal to the posterior middle temporal sulcus (PMTS). In M2, we recorded just above the PMTS.

Quantification and statistical analyses

Behavioral analysis: Free-viewing

The viewing times for the VTA-EM paired and unpaired fractal were compared between the pre-and post-VTA-EM free-viewing sessions. The viewing time was computed as the time that the subject’s eye position was inside a 6° window centered on the fractal. Subsequently, a preference index was computed based on viewing times per fractal condition with the following equation:

preferenceindex(PI)=TimeSTIM−TimeCONTROLTimeSTIM+TimeCONTROL

The preference index (PI) can vary between -1 and 1 with positive PI’s indicating that the monkey spent more time looking at the VTA-EM paired fractal compared to the non-paired fractal and negative PI’s indicating that the monkey spent more time looking at the unpaired fractal compared to the VTA-EM paired fractal. A PI of 0 indicates no preference for either fractal.

We compared the PI between the pre-and post-VTA-EM free-viewing sessions by subtracting the PI of the pre-VTA-EM free-viewing sessions from the PI of the post-VTA-EM free-viewing session. Outcomes from this subtraction that are positive represent relatively longer viewing times for the VTA-EM paired fractal during the post-VTA-EM session compared to the pre-VTA-EM session. We tested for each monkey the statistical significance of the difference in PI between pre-and post-VTA-EM with a Wilcoxon signed-rank test.

Electrophysiology analysis: Fixation task

For unaborted trials, spikes were counted in time windows of [-200ms 0ms] for baseline and [50ms 250ms] for the gross response, where 0 is stimulus onset. A split-plot ANOVA was used to select responsive units, with as a repeated factor baseline versus gross activity and as a between-trial factor the fractal stimuli. Units were included for analysis if all 6 (M1) or 7 (M2) recording epochs produced a significant outcome, at a significance level of p<0.05, for either the repeated factor or the interaction effect of the split-plot ANOVA. All analysis was performed on net responses (gross – baseline) unless specified otherwise.

To assess the effect of pairing VTA-EM with a fractal image compared to a non-paired fractal image, we first calculated the mean response per epoch and fractal condition. We subtracted the response to the CON fractal from the response to the STIM fractal (STIM-CON) and the response to the DUM1 fractal from the response to the DUM2 fractal (DUM1-DUM2) for every recording epoch. An orthogonal regression line (Carr, 2021) was fitted to the STIM-CON and DUM1-DUM2 data over the pre-VTA-EM epoch vs. each post-VTA-EM epoch. A 99% confidence interval (i.e. Bonferroni corrected 95% confidence interval for 1 pre-VTA-EM epoch vs. 5 post-VTA-EM epochs) was determined by bootstrapping (1000 resamplings). The Pearson correlation coefficient (r) was computed. We employed orthogonal regression because both variables were random variables and avoid regression to the mean effects. Ordinary linear regression and robust linear regression produced highly similar significant effects.

This regression analysis was also applied to the gross spiking data that was divided into time windows of 50ms from [-150ms 300ms], 0 being stimulus onset.

Electrophysiology analysis: Receptive field mapping

Receptive field mapping sessions were analyzed as described in (Herpers et al., 2021), but for net responses instead of normalized responses. For unaborted trials, spikes were counted in time windows of [-200ms 0ms] for baseline and [50ms 250ms] for the gross response, where 0 is stimulus onset. To be selected for analysis, channels needed 1) to produce a significant response in all 6 (M1) or 7 (M2) fixation epochs. 2) produce a significant response in the receptive field mapping task, based on a split-plot ANOVA, with as a repeated factor baseline versus gross activity and the between-trial factor of checkerboard position. The significance level of either the repeated factor or the interaction effect had to be p<0.05 to meet the second criterium to be selected for analysis. Average receptive field maps are presented together with the location of the fractal for each monkey in Figure S1.

Data and code availability

The behavioral and electrophysiological datasets are available at Mendeley Data: https://data.mendeley.com/v1/datasets/6xkbhhwbxn/draft?preview=1. Custom scripts to generate the figures and statistical outcomes in this study are available in the same online repository. Any additional information and/or raw data can be requested from the senior author upon request.

Results

We investigated in two macaque monkeys (M1 and M2) whether a brief period during which fractals were paired with VTA-EM (60 pairings) affects neural responses in PIT. To perform VTA-EM, microwire brush electrodes (Bondar et al., 2009) were implanted chronically in the right hemisphere of monkey M1 and the left hemisphere of monkey M2 (Arsenault et al., 2014). During the VTA-EM pairing epochs, paired and unpaired stimuli were presented peripherally during passive fixation. Before we started the recording experiments, we verified that the implanted VTA electrodes affected free-viewing behavior for fractal stimuli presented in the peripheral visual field in a stimulus-VTA-EM association paradigm (Arsenault et al., 2014; Herpers et al., 2021).

Free-viewing and VTA-EM stimulus pairing

In previous studies, we showed that pairing VTA-EM with a visual stimulus presented in the center of the visual field produces a behavioral preference for the paired stimulus (Arsenault et al., 2014; Herpers et al., 2021). To verify whether pairing VTA-EM with a visual stimulus presented in the peripheral visual field could produce a behavioral preference as well, we employed the same protocol as in Herpers et al. (2021). However, during VTA-EM-stimulus pairing, fractals were presented at 3° eccentricity instead of centrally (Figure 1B). We chose this stimulus position since the main experiment of this study also involved the pairing of VTA-EM with a fractal at 3° eccentricity.

The VTA-EM-stimulus pairing was performed during a passive fixation task (Figure 1B). In the task, two fractals (size = 4°) were presented in pseudo-random order, at 3° eccentricity in the lower visual field contralateral to the stimulated hemisphere. A trial consisted of the presentation of one of two fractals and the presentation frequency of the two fractals was equated. One of the two fractals was paired with VTA-EM (~60 VTA-EM stimulus pairing events, see Methods for EM parameters).

Before, and following a VTA-EM pairing session, the monkeys performed a free-viewing task (Figure 1B). This task was used to measure whether there was a shift in preference after VTA-EM for either fractal. The preference for each fractal was measured by the amount of time spent viewing the fractals. In this task, the same two fractals that were used during VTA-EM stimulus pairing were presented simultaneously. The monkey received a juice reward whenever its gaze was within the limits of the screen, while the fractals were presented to the left (eccentricity of 3°) and right of the middle of the screen on the horizontal meridian. The fractals interchanged positions randomly over trials.

This procedure of two free-viewing epochs enveloping a VTA-EM epoch was performed with 15 (M1) and 13 (M2) arbitrarily chosen fractal pairs. To quantify the viewing times for every free-viewing epoch, a Preference Index (PI) for viewing times was computed by subtracting the time looked at the non-paired fractal from the time looked at the VTA-EM paired fractal, divided by the sum of both viewing times. PI can range from -1 to 1, with a positive PI corresponding to a longer viewing time for the VTA-EM paired fractal compared to the non-paired fractal and a negative PI corresponding to relatively more viewing time for the unpaired fractal. A PI of 0 indicates no preference for either fractal.

We subtracted the PI of the pre-VTA-EM session from the PI of the post-VTA-EM session (post_VTA_EM PI – pre_VTA_EM PI; Figure 1C). Positive outcomes from this subtraction indicate longer viewing times for the VTA-EM paired fractal after VTA-EM compared to before VTA-EM. This computation showed on average relatively longer viewing times for the VTA-EM paired fractal post-VTA-EM compared to before VTA-EM pairing in both monkeys (Figure 1C, one-sided Wilcoxon signed-rank test, M1: p = 0.015, M2: p = 0.001). These outcomes are in line with the findings of (Herpers et al., 2021), where VTA-EM paired with a fractal in the central visual field increased viewing times for that fractal. When presenting the VTA-EM paired fractal in the peripheral visual field, the behavioral stimulus preference of the monkey can be changed as well, although the effect is numerically smaller than when the fractal is presented foveally during VTA-EM (see Figure 1C and D in Herpers et al. (2021)).

Recordings in PIT interleaved with stimulus-VTA-EM pairing

After finding that pairing a peripheral fractal with VTA-EM during a short epoch (60 VTA-EM pairing events) has a behavioral effect on viewing preference, we investigated whether pairing a fractal with VTA-EM for the same period could also affect neuronal responses in higher-order visual cortex. To this end, we measured the MUA of PIT neurons in response to visual stimuli before and after VTA-EM-stimulus pairing. Unlike in previous imaging and recordings studies (Arsenault & Vanduffel, 2019; Herpers et al., 2021; Murris et al., 2021), but as for the behavioral free-viewing test (see above), the visual stimulus was paired with VTA-EM only for brief periods of ~60 VTA-EM pairing trials during a single daily session. In a single daily session (Figure 2B), 5 (M1) or 6 (M2) VTA-EM pairing epochs were performed with the same paired and unpaired fractals, interleaved with MUA recording epochs with four fractals, including the paired and unpaired fractal. The recorded hemisphere was ipsilateral to the VTA electrode and fractals were presented in the visual field contralateral to this hemisphere. During the MUA recording epochs, the four fractals were presented in the lower visual field at 3° eccentricity in pseudo-random order during passive fixation. During the VTA-EM epochs, two of these four fractals were presented, one on each trial, at the same eccentricity as during the recordings. Note that this visual field position was identical to the position employed in the VTA-EM epochs enveloped by free-viewing epochs for which a behavioral effect of VTA-EM fractal pairing was observed using the same number of VTA-EM stimulus pairings. During the VTA-EM epoch, one of the fractals was paired with VTA-EM (STIM) and the other fractal was not paired and served as a control (CON). The two other fractals presented during the recording epochs were not included in the VTA-EM epochs and are considered as “dummy” fractals (DUM1 and DUM2). They served to assess the impact of exposure per se during the VTA-EM epoch and to enhance the number of stimuli during the recording epoch, thereby reducing adaptation effects. We recorded MUA with a multi-channel Plexon V-Probe in PIT, in sites where the receptive fields overlapped with the fractal’s visual field position (Figure S1). The VTA-EM pairing and recording protocol was repeated for 15 (M1) and 20 (M2) daily sessions, with 4 novel randomly chosen fractals in each daily session.

For the analysis of the MUA in response to the fractals, we first employed a split-plot ANOVA per channel and per recording epoch to determine if the MUA produced a significant response (see Methods). Recorded channels were selected for analysis if the response to any of the four fractals proved to be significant across all 6 (M1) or 7 (M2) recording epochs. For the selected channels, the average net (baseline subtracted) response for the 4 fractal conditions was computed (Figure 3). The average response rate decreased for all 4 conditions over recording epochs in both monkeys. Possible explanations for the decrease in response with epoch number include repetition suppression (Desimone, 1996; Vogels, 2016), or a deterioration in the quality of the recorded electrophysiological signal over time.

Figure 3. Mean response to the different fractal conditions.

Figure 3

Open in a new tab

Recordings in PIT were interleaved with VTA-EM epochs. Fractal conditions are either paired with VTA-EM (STIM), not paired with VTA-EM (CON), or not presented during the VTA-EM epoch (DUM1 & DUM2). In both monkeys, the mean response to the fractals decreases over recording epochs. The greatest decrease in response can be observed for the CON fractal after the first VTA-EM epoch.

To ensure a stable electrophysiological signal, we waited 1 hour between lowering the multi-channel probe in the cortex and starting the daily recording session. However, as a daily session could last almost 3 hours, it is not unlikely that the quality of the signal could fluctuate over this period. To test whether there was a deterioration of the recorded signal, we computed the signal-to-noise ratio (SNR) of the signal by dividing the variance of the amplitude of the sorted spikes (signal) by the variance of the amplitude of the signal in between spikes (noise). We compared the SNR of the first epoch with the SNR of the second, and with the SNR of the last epoch. In M1, the SNR increased significantly (change in SNR = 7.3%) when comparing recording epoch 1 and epoch 2 (Wilcoxon signed-rank test, p = 0.0069, Bonferroni corrected for 2 comparisons) and there was a significant decrease in SNR when comparing epoch 1 and epoch 6 (-5.2%; Wilcoxon signed-rank test, p = 0.0002, Bonferroni corrected for 2 comparisons). In M2, there was a non-significant decrease in SNR between epoch 1 and epoch 2 (-3.7%; Wilcoxon signed-rank test, p = 0.17 Bonferroni corrected for 2 comparisons) and a significant decrease in SNR between epoch 1 and epoch 7 (-10.5%; Wilcoxon signed-rank test, p = 0.00001, Bonferroni corrected for 2 comparisons). This suggests that the decrease in spike rate between the first and the last epoch could be explained, in part, by a deterioration of the SNR of the recorded signal. However, the decrease in spike rate between the first and second epoch cannot be explained by a change in the SNR of the recorded signal, as the SNR increased (M1) or stayed the same (M2) for these two recording epochs.

In the pre-VTA-EM recording epoch, the average response was higher for the CON than for the STIM images, even though the CON and STIM images were selected at random from the effective images of one channel out of 16 (see Methods). At first sight, this makes the stronger average reduction in response for the CON compared to STIM images with VTA-EM pairing difficult to interpret. However, individual units differed markedly in their difference in response between the STIM and CON images in the pre-VTA-EM recording epoch, with some units showing no difference in response between the STIM and CON images. This allowed us to assess whether VTA-EM stimulus pairing affected the response differences between STIM and CON fractals. We quantified the units’ difference in response for the STIM and CON fractal by subtracting the response to the CON fractal from the response to the STIM fractal (STIM-CON) for every recording epoch. For every post-VTA-EM epoch, we plotted the STIM-CON values per unit of the pre-VTA-EM epoch against the STIM-CON values of the same unit in the post-VTA-EM epochs. Then, to relate the pre-post-VTA-EM response differences, we fitted an orthogonal linear regression line and calculated the Bonferroni corrected 95% confidence interval of the intercept for 5 regression line fittings by bootstrapping (Figure 4). A positive intercept of the fitted regression line (i.e. mean post-VTA-EM response difference when there was no difference in response between the STIM and CON fractal in the pre-VTA-EM epoch) indicates that VTA-EM pairing increased the response to the STIM relative to the CON fractal. In each monkey, the intercept was significantly positive when correlating the STIM-CON response differences of the first and second recording epoch (Figure 4), indicating that 60 VTA-EM stimulus pairings were sufficient to induce a relative increase of the response to the paired fractal. The relative increase in response was numerically small, about 3 spikes/s, but consistent for the two monkeys. Similar relative increases in response for the paired fractal were observed for subsequent VTA-EM stimulus pairing blocks of a daily session (Figure 4): the intercept of the fitted regression was significantly greater than 0 for all post-VTA-EM epochs (Figure 4; see Figure S2 for individual monkey data). Unexpectedly, the difference in response between STIM and CON fractals did not increase with the number of VTA-EM epochs. This was consistent across animals (Figure S2).

Figure 4. Orthogonal linear regression analysis of pre-VTA-EM recording epoch vs. post-VTA-EM epochs.

Figure 4

Open in a new tab

Relative responses were calculated between the STIM-CON fractal pair and the DUM1-DUM2 fractal pair pre-and post-VTA-EM. Unit responses for the pre-VTA-EM recording epoch were plotted against all post-VTA-EM epochs separately and an orthogonal regression line was fitted. A 95% (Bonferroni corrected) confidence interval was computed by bootstrapping, indicated by cyan outline. In the left section, the first recording epoch for the STIM-CON fractal pair and the DUM1-DUM2 fractal pair is deconstructed into the separate data per monkey. Results were consistent over monkeys, and hence, we pooled the data for analysis. In the right graph, the intercept (±95% Bonferroni corrected confidence intervals corrected for 5 regression line fittings) are plotted per recording epoch pooled over monkeys. The intercept of the regression line is significantly greater than 0 for all epochs for the STIM-CON fractal pair. However, the effect did not increase after the initial VTA-EM epoch. This effect was not observed when we performed the same analysis on the dummy fractal pair.

We performed the same regression analysis for the dummy fractal pair (DUM1-DUM2). None of the intercepts for the DUM1-DUM2 fractals were significantly different from 0. The slopes of the regression lines (Figure S3) for the dummy fractal response differences tended to be smaller than 1. This followed from a reduction of stimulus selectivity, as the variance of the absolute response difference decreased during the course of testing. Except for the later recording epochs, no such trend for slopes smaller than 1 was present for the STIM-CON responses differences (Figure S3). In sum, we observed a relative increase in response for the STIM fractal compared to the CON fractal after VTA-EM-stimulus pairing. This change in response occurred already after one VTA-EM epoch but did not increase further when repeating the VTA-EM stimulus pairing in subsequent epochs. The change in response was specific for the fractals presented during the VTA-EM stimulus pairing epochs since it was not observed for the dummy fractals, that were not presented during the VTA-EM epoch.

The latency of the relatively greater response to the STIM fractal compared to the CON fractal after VTA-EM stimulus pairing revealed interesting additional information. When this effect occurs early in the response, it could point to an increase in spike rate that is caused by increased bottom-up activity feeding into the area, or that emerges locally within PIT from ascending dopaminergic fibers. When the effect would occur late in the response, it could point towards an increase in spike rate as a result of top-down influences. To examine the latency of the effect found in the previous regression analysis, we performed the same orthogonal regression fitting for short windows of 50ms, every 50ms, starting from 150ms before stimulus onset until 300ms after stimulus onset. We computed a Bonferroni corrected (for 5 time-bins from 50ms-300ms after stimulus onset, which are the bins in which a response can be expected in PIT) 95% confidence interval of the intercept by bootstrapping. We performed this analysis only for the pre-VTA-EM recording epoch vs. the first post-VTA-EM recording epoch since the effect already occurred after the first VTA-EM epoch and did not increase after subsequent VTA-EM epochs. There was no significant difference for any 50ms windows for the DUM1-DUM2 control fractals when pooling data over monkeys (Figure 5). For the STIM-CON fractal pair, the intercept was significantly different from 0 from the 50ms-100ms window until the 250ms-250ms window (see Figure S4 for correlation coefficients between the response differences in pre-and post-VTA-EM epoch recordings for the different analysis windows). Hence, the latency of the effect occurred early after stimulus onset. For monkey M1 and M2, the intercept was significantly different from 0 in the windows 100ms-300ms and 50ms-250ms, respectively. This analysis using short windows suggests that the latency of the relatively increased response for the STIM fractal compared to the CON fractal is unlikely to be the effect of top-down influences, as the effect occurs early in the response.

Figure 5. Time window regression analysis.

Figure 5

Open in a new tab

Per window of 50ms, for every 50ms, relative responses were calculated between the STIM-CON fractal pair and the DUM1-DUM2 fractal pair of the first and second recording epoch. We fitted an orthogonal regression line per time window and plotted the intercept together with the 95% bootstrapped confidence interval (Bonferroni corrected for 5 time-bins 50ms-300ms after stimulus onset). Data points are plotted at the beginning of the 50ms time window. The left panel shows the sliding window regression analysis pooled over monkeys. From 50ms-250ms, the intercept was significantly greater than 0 for the STIM-CON fractal pair, whereas no significant outcome was observed for the DUM1-DUM2 fractal pair. This shows that the latency of the relative increase in firing rate is short. The right panel shows the same analysis for each individual monkey, where for M1, STIM-CON was significantly different from 0 from 100ms-300ms, and for M2, STIM-CON was significantly different from 0 in time windows 50ms-250ms.

During the data analysis, we noted that, especially in M2, the CON fractals were presented more frequently than the STIM fractals during VTA-EM-stimulus pairing. Indeed, during a VTA-EM epoch, the CON fractal could have been presented more often if the monkey aborted more CON than STIM trials during the presentation of those fractals since only completed trials were used for trial equation over conditions. A higher frequency of presentations of a particular stimulus could lead to a greater neuronal adaptation for that stimulus. To address this potential confound, we computed two indices: the fractal presentation index (FPI) representing the relative number of fractal presentations (including aborted trials) during the first VTA-EM epoch for the STIM and CON conditions, and the STIM-CON index (SCI) representing the response difference between the STIM and CON fractal over the pre-and post-VTA-EM epochs. The indices were computed as follows:

Fractalpresentationindex(FPI)=npresentationsSTIM−npresentationsCONnpresentationsSTIM+npresentationsCON
STIM−CONindex(SCI)=STIMpost-EM−STIMpre-EMISTIMpost-EMI+ISTIMpre-EMI−CONpost-EM−CONpre-EMICONpost-EMI+ICONpre-EMI

With STIM pre-EM and CON pre-EM being the net response to the STIM and CON stimulus in the first (pre-VTA-EM) recording epoch, respectively, and STIM post-EM and CON post-EM being the net response to the STIM and CON stimulus in the second (post-VTA-EM) epoch, respectively. In agreement with the linear regression analysis of Figure 4, the median SCI was significantly greater than 0 (Wilcoxon signed-rank test; M1: p = 9x10-7, M2: p = 3x10-5), indicating a relatively greater response to the STIM fractal after VTA-EM stimulus pairing.

Then, for each unit, we plotted the PI against the SCI (Figure 6) and we computed the Spearman correlation between the two indices. If the response difference between the first and second recording epoch for the CON and STIM trials during the recordings resulted from a difference in stimulus presentation frequency and resulting adaptation during the VTA-EM epoch, one would expect a positive correlation between the two indices. However, for both monkeys, there was no significant correlation between the number of stimulus presentations during the first VTA-EM epoch and the relative response to the STIM and CON fractal pre-and post-VTA-EM (Spearman rank correlation; M1: ρ = -0.09, p = 0.24; M2, ρ = -0.01, p = 0.84). We repeated this analysis including only the last 50, 20, or 10 trials of the VTA-EM epoch, since the last trials may contribute most to a putative repetition suppression effect in the next recording epoch. However, we found no correlation between the relative number of fractal presentations and the relative response difference between the STIM and CON fractal pre-and-post VTA-EM in any of these trial selections. These analyses show that the difference in response rate between the first and second recording epoch did not depend on the frequency of fractal presentations during the first VTA-EM epoch.

Figure 6. The effect of the difference in fractal presentations during VTA-EM on the difference in relative firing rate between pre-and post-VTA-EM and fractal conditions.

Figure 6

Open in a new tab

During the VTA-EM epochs, trials aborted during stimulus presentation were not used for equating trial numbers over conditions. Hence, when the monkey would make more aborts for one condition compared to the other condition, this could result in more repetition suppression for the condition with more presentations. A fractal presentation index (FPI) was calculated for the relative number of fractal presentations, and a STIM-CON index (SCI) was calculated for the relative difference in response between the STIM and CON fractal pre- and post-VTA-EM (see Results for equations). The red line corresponds to the least-squares line. The Spearman correlation (ρ) between the two indices was not significant for either monkey (Spearman rank, M1: p = 0.24; M2, p = 0.84). This suggests that the difference in response between the STIM-CON fractals between pre-and post-VTA-EM is not dependent on the number of stimulus presentations during the VTA-EM epoch.

Discussion

Long periods of VTA-EM-stimulus pairings induce an increase in the cortical representation for that stimulus (Arsenault & Vanduffel, 2019; Bao et al., 2001; Herpers et al., 2021; Murris et al., 2021). We showed by interleaving electrophysiological recordings in PIT with epochs of VTA-EM, that after one short VTA-EM-stimulus pairing epoch (~60 pairing events), multi-unit activity increased for the stimulus paired with VTA-EM (STIM) compared to the fractal that was not paired with VTA-EM (CON). We found that the relative increase in firing rate for the paired stimulus already occurred early in the response. Interestingly, neural responses did not further increase for the STIM fractal after subsequent VTA-EM pairing epochs within a daily session.

In agreement with previous studies (Herpers et al., 2021), VTA-EM-stimulus pairing in the free-viewing paradigm resulted in an increased inspection of the stimulus paired with the reward. This behavioral result is also in line with Arsenault et al., (2014), showing that VTA-EM-stimulus pairings increased the monkey’s preference for a paired stimulus (a shape) using both classical and operant conditioning paradigms. The effect of behavioral preference appears to be smaller in the present study compared to previous studies, likely because of the peripheral location of the stimuli during VTA-EM stimulus pairing. Here we showed that the same number of VTA-EM pairings with a peripheral stimulus of the same class (fractals) increased the visual cortical response to the paired stimulus. We do not know whether the increased PIT response contributes causally to the behavioral preference in the free-viewing task. A potential link between behavior and neural responses can be made based on seminal work by Hikosaka and colleagues (2019), who have mapped out a basal ganglia circuit that is involved in the control of eye movements based on reward-induced stimulus value (Hikosaka et al., 2019). This circuit receives dopaminergic input that signals the value and/or saliency of the stimulus during learning. Visual cortex, including PIT (Kravitz et al., 2013), provides input to this basal ganglia circuit. Thus, by combining information from the VTA and PIT in the basal ganglia, an increased visual cortical response to the paired stimulus could in principle contribute to the behavioral preference shift following VTA-EM stimulus pairing.

Unlike our previous studies (Arsenault & Vanduffel, 2019; Herpers et al., 2021; Murris et al., 2021), we did not control for attention when measuring the responses to VTA-EM paired and unpaired images. However, the fast onset of the increase in response to the paired fractal makes it unlikely that the increase in response results from top-down, attentional signals. A more likely explanation is that the increased response originates locally or is inherited from earlier visual cortical areas. On the one hand, pairing a stimulus with activation of the VTA might lead to an increased neuronal response through long-term potentiation (LTP). The VTA has dopaminergic projections throughout the cortex, including temporal cortex, and especially pronounced in layer 1 (Zubair et al., 2021). Dopamine could facilitate spike-timing-dependent plasticity and increase the excitatory postsynaptic potentials of cells responsive to the paired stimulus within minutes (Bissieére et al., 2003; Xu & Yao, 2010). Thus, the increased responses might be a result of local neural plasticity mechanisms driven by direct inputs from the VTA. On the other hand, functional imaging and electrophysiological studies have shown changes in early visual cortical areas after albeit long-term pairing of the visual stimulus and a reward (Arsenault et al., 2013; Frankó et al., 2010; Ghazizadeh et al., 2018, 2020; Kaskan et al., 2017) or VTA-EM (Murris et al., 2021). Moreover, although PIT showed the strongest effects, our previous studies documenting cortical plasticity at the fMRI level induced by VTA-EM-stimulus pairings also revealed changes in early visual areas (Arsenault & Vanduffel, 2019; Murris et al., 2021). Hence, the response changes we observed here at the neuronal level in PIT might, at least partially, be inherited from early visual cortical areas.

We expected an increase in the response to the paired stimulus with an increasing number of VTA-EM epochs. However, after an initial relative increase in response to the paired stimulus compared to the unpaired stimulus after the first VTA-EM stimulus pairing epoch, there was no subsequent increase in response following subsequent VTA-EM epochs. One potential explanation of the lack of increase of the effect after the second and subsequent VTA-EM epochs might be habituation of the VTA-EM response. VTA dopaminergic neurons show a strong depression of the response after stimulation. The first pulse of axonal dopamine release depletes already 60% of the available dopamine (Liu et al., 2021). Replenishing the stores of releasable dopamine may be slow, as the depression of dopamine release can last for tens of seconds after a single action potential (Liu et al., 2021). To take these factors into account, we designed the VTA-EM procedure so that the average interval between two VTA-EM events was 17.4s to avoid habituation of the VTA-EM response and the interval between VTA-EM epochs was ~5 minutes. Nonetheless, these intervals coupled with the high stimulation frequency might have been insufficient to completely overcome the habituation of the dopaminergic response. In contrast, when performing multiple free-viewing tests with different, novel fractals in single a day with the same VTA-EM stimulation parameters and duration as in the electrophysiological sessions, we observed that a behavioral effect of VTA-EM was still present for the later tests. This indicates that there was sufficient VTA activity after the first VTA-EM epoch in subsequent epochs to produce a behavioral effect for novel stimuli. However, a difference from the electrophysiology sessions was that the interval between successive VTA-EM epochs in the free-viewing sessions was ~20 minutes and only 5 minutes in the electrophysiology sessions. Therefore, longer intervals between VTA-EM epochs might give more time to replenish the releasable dopamine between VTA-EM epochs, which could lead to an increasing effect over time.

As neural activity, visual performance can also be enhanced by VTA-EM. For example, pairing a weak, peri-threshold visual stimulus with VTA-EM for several days increases the perceptual sensitivity for the paired stimulus (Arsenault & Vanduffel, 2019; Murris et al., 2021). This perceptual change has been related to task-irrelevant perceptual learning in which perceptual thresholds increase after stimulus-reward pairing, even for subliminal stimuli (Seitz et al., 2009; Seitz & Watanabe, 2005; Watanabe & Sasaki, 2015). We do not know whether our brief epoch of VTA-EM stimulus pairing has perceptual consequences. If there would be any direct changes in perception after VTA-EM, it is expected that changes would occur already after the first VTA-EM epoch, as the relative increase in response in PIT occurred after only 60 VTA-EM events. Yet the postulated change in perception would only be small, as the relative response rate only increased by 3 spikes/second. Note that we employed suprathreshold stimuli, which at least in adults do not produce task-irrelevant perceptual learning (Frank et al., 2021; Tsushima et al., 2008). Previous studies that showed perceptual effects of VTA-EM stimulus pairing employed peri-threshold stimuli and these were presented during the execution of a demanding task that drew away attention from the paired stimulus (Arsenault & Vanduffel, 2019; Murris et al., 2021). It remains to be seen whether a brief VTA-EM stimulus pairing epoch is sufficient to produce perceptual and neural changes when using such weak visual stimuli under orthogonal task conditions, or whether sleep is required to observe these changes in perception and neural activity (Tamaki et al., 2020).

Overall, responses decreased over recording epochs for all stimulus conditions. Our analyses suggested that the decrease in response is not because of a reduced SNR of the recorded signal between epochs, at least not for the first and second epoch for which the decrease in response strength was the largest. Thus, it is tempting to explain at least part of the decreased responses as a result of repetition suppression, which is a decrease in neural response when repeating the same stimulus, a well-documented phenomenon in macaque IT (Desimone, 1996; Vogels, 2016). The numerically small decreases in responses with repetition, except for the CON stimulus, might be a result of the long intervals between repetitions of the same stimulus, the presence of intervening stimuli, and because we compare blocks with many repetitions. Indeed, repetition suppression is typically the largest at the first repetition (Vogels, 2016), although this may be not true for long interstimulus intervals, as in the present study. The decrease in response across epochs also tended to be smaller for the dummy images compared to the CON image, which might be because the former stimuli were not presented during the VTA-EM pairing epochs and thus more time elapsed between the presentation of these stimuli in subsequent recording epochs.

A limitation of the current study is the lack of neuronal selectivity by electrical stimulation, as non-dopaminergic VTA neurons can be activated as well. Further studies employing neuron-type specific stimulation, e.g. using optogenetics, or pharmacological manipulations with dopamine receptor antagonists will be necessary to assess the definitive role of dopamine in the observed response increases.

Our study shows that a brief epoch of pairing of a visual stimulus with VTA-EM increases the responses of visual cortical neurons to the paired stimulus. The response change following VTA-EM stimulus pairing occurred early in the course of the response. A recent study showed that a brief pairing of a visual stimulus with an aversive stimulus increases the response of primary visual cortical neurons (Li et al., 2019). This response increase had an early response onset, suggesting a local or bottom-up mechanism. Moreover, the pairing of a reward with a stimulus changes fMRI activity in the representation of the paired stimulus in macaque early visual cortex (Arsenault et al., 2013). Both the current study and Li et al. (2019) suggest that the association of a visual stimulus and a highly salient event, such as an aversive stimulus or VTA-EM, potentiates the visual cortical response to the predictive stimulus.

Supplementary Material

Supplementary Material

Acknowledgments

The authors thank J.T. Arsenault, M. Depaep, W. Depuydt, C. Fransen, J. Helin, A. Hermans, P. Kayenbergh, G. Meulemans, I. Puttemans, C. Ulens, M. Verbeeck and S. Verstraeten for technical and administrative support. This project received funding from KU Leuven (grants C14/17/109 and C14/21/111); Fonds Wetenschappelijk Onderzoek- Vlaanderen (FWO-Flanders) (grants G0D5817N, G0B8617N, G090714N, G0E0520N, G088418N, and Odysseus G0007.12); the European Union’s Horizon 2020 Framework Programme for Research and Innovation under grant agreement 945539 (Human Brain Project SGA3).

Footnotes

Author Contributions

Conceptualization: R.V., W.V. & J.H. Methodology: R.V., W.V. & J.H. performed research: J.H. Analysis: J.H. Visualization: J.H. Writing - original draft: J.H. Writing - review and editing: R.V. & W.V. Funding acquisition: R.V., and W.V. Resources and supervision: W.V. and R.V.

Declaration Of Interest

The authors declare no competing interests.

References

  1. Arsenault JT, Nelissen K, Jarraya B, Vanduffel W. Dopaminergic reward signals selectively decrease fMRI activity in primate visual cortex. Neuron. 2013;77(6):1174–1186. doi: 10.1016/j.neuron.2013.01.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Arsenault JT, Rima S, Stemmann H, Vanduffel W. Role of the primate ventral tegmental area in reinforcement and motivation. Current Biology. 2014;24(12):1347–1353. doi: 10.1016/j.cub.2014.04.044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Arsenault JT, Vanduffel W. Ventral midbrain stimulation induces perceptual learning and cortical plasticity in primates. Nature Communications. 2019;10(1):1–13. doi: 10.1038/s41467-019-11527-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bao S, Chan VT, Merzenich MM. Cortical remodelling induced by activity of ventral tegmental dopamine neurons. Nature. 2001;412(6842):79–83. doi: 10.1038/35083586. [DOI] [PubMed] [Google Scholar]
  5. Berke JD. What does dopamine mean? Nature Neuroscience. 2018;21(6):787–793. doi: 10.1038/s41593-018-0152-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Berridge KC. The debate over dopamine’s role in reward: The case for incentive salience. Psychopharmacology. 2007;191(3):391–431. doi: 10.1007/s00213-006-0578-x. [DOI] [PubMed] [Google Scholar]
  7. Berridge KC, Kringelbach ML. Affective neuroscience of pleasure: Reward in humans and animals. Psychopharmacology. 2008;199(3):457–480. doi: 10.1007/s00213-008-1099-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bissieére S, Humeau Y, Luüthi A. Dopamine gates LTP induction in lateral amygdala by suppressing feedforward inhibition. Nature Neuroscience. 2003;6(6):587–592. doi: 10.1038/nn1058. [DOI] [PubMed] [Google Scholar]
  9. Bondar Iv, Leopold DA, Richmond BJ, Victor JD, Logothetis NK. Long-term stability of visual pattern selective responses of monkey temporal lobe neurons. PLoS ONE. 2009;4(12) doi: 10.1371/journal.pone.0008222. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bromberg-Martin ES, Matsumoto M, Hikosaka O. Dopamine in Motivational Control: Rewarding, Aversive, and Alerting. Neuron. 2010;68(5):815–834. doi: 10.1016/j.neuron.2010.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Buonomano DV, Merzenich MM. Cortical plasticity: From synapses to maps. Annual Review of Neuroscience. 1998;21:149–186. doi: 10.1146/annurev.neuro.21.1.149. [DOI] [PubMed] [Google Scholar]
  12. de Jong JW, Afjei SA, Pollak Dorocic I, Peck JR, Liu C, Kim CK, Tian L, Deisseroth K, Lammel S. A Neural Circuit Mechanism for Encoding Aversive Stimuli in the Mesolimbic Dopamine System. Neuron. 2019;101(1):133–151.:e7. doi: 10.1016/j.neuron.2018.11.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Desimone R. Neural mechanisms for visual memory and their role in attention. Proceedings of the National Academy of Sciences. 1996;93(24):13494–13499. doi: 10.1073/pnas.93.24.13494. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Draganski B, Gaser C, Busch V, Schuierer G, Bogdahn U, May A. Changes in grey matter induced by training. Nature. 2004;427:311–312. doi: 10.1038/427311a. [DOI] [PubMed] [Google Scholar]
  15. Engelhard B, Finkelstein J, Cox J, Fleming W, Jang HJ, Ornelas S, Koay SA, Thiberge SY, Daw ND, Tank DW, Witten IB. Specialized coding of sensory, motor and cognitive variables in VTA dopamine neurons. Nature. 2019;570(7762):509–513. doi: 10.1038/s41586-019-1261-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Carr F. Orthogonal Linear Regression. MATLAB Central File Exchange; 2021. Sep 24, https://Nl.Mathworks.Com/Matlabcentral/Fileexchange/16800-Orthogonal-Linear-Regression. [Google Scholar]
  17. Frank SM, Bründl S, Frank UI, Sasaki Y, Greenlee MW, Watanabe T. Fundamental differences in visual perceptual learning between children and adults. Current Biology. 2021;31(2):427–432.:e5. doi: 10.1016/j.cub.2020.10.047. [DOI] [PubMed] [Google Scholar]
  18. Frankó E, Seitz AR, Vogels R. Dissociable neural effects of long-term stimulusreward pairing in macaque visual cortex. Journal of Cognitive Neuroscience. 2010;22(7):1425–1439. doi: 10.1162/jocn.2009.21288. [DOI] [PubMed] [Google Scholar]
  19. Ghazizadeh A, Fakharian MA, Amini A, Griggs W, Leopold DA, Hikosaka O. Brain Networks Sensitive to Object Novelty, Value, and Their Combination. Cerebral Cortex Communications. 2020;1(1) doi: 10.1093/texcom/tgaa034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Ghazizadeh A, Griggs W, Leopold DA, Hikosaka O. Temporal-prefrontal cortical network for discrimination of valuable objects in long-term memory. Proceedings of the National Academy of Sciences of the United States of America. 2018;115(9):E2135–E2144. doi: 10.1073/pnas.1707695115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Gilbert CD, Li W. Adult Visual Cortical Plasticity. Neuron. 2012;75(2):250–264. doi: 10.1016/j.neuron.2012.06.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Glimcher PW. Understanding dopamine and reinforcement learning: The dopamine reward prediction error hypothesis. 2011;108(42) doi: 10.1073/pnas.1115170108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hamid AA, Pettibone JR, Mabrouk OS, Hetrick VL, Schmidt R, vander Weele CM, Kennedy RT, Aragona BJ, Berke JD. Mesolimbic dopamine signals the value of work. Nature Neuroscience. 2015;19(1):117–126. doi: 10.1038/nn.4173. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Herpers J, Arsenault JT, Vanduffel W, Vogels R. Stimulation of the ventral tegmental area induces visual cortical plasticity at the neuronal level. Cell Reports. 2021;37(6):109998. doi: 10.1016/j.celrep.2021.109998. [DOI] [PubMed] [Google Scholar]
  25. Hikosaka O, Yasuda M, Nakamura K, Isoda M, Kim HF, Terao Y, Amita H, Maeda K, Wurtz RH. Multiple neuronal circuits for variable object-action choices based on short-and long-term memories. PNAS. 2019;116(52):26313–26320. doi: 10.1073/pnas.1902283116/-/DCSupplemental. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Hughes RN, Bakhurin KI, Petter EA, Watson GDR, Kim N, Friedman AD, Yin HH. Ventral Tegmental Dopamine Neurons Control the Impulse Vector during Motivated Behavior. Current Biology. 2020;30(14):2681–2694.:e5. doi: 10.1016/j.cub.2020.05.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jay TM. Dopamine : a potential substrate for synaptic plasticity and memory mechanisms. 2003;69:375–390. doi: 10.1016/S0301-0082(03)00085-6. [DOI] [PubMed] [Google Scholar]
  28. Kaskan PM, Costa VD, Eaton HP, Zemskova JA, Mitz AR, Leopold DA, Ungerleider LG, Murray EA. Learned Value Shapes Responses to Objects in Frontal and Ventral Stream Networks in Macaque Monkeys. Cerebral Cortex. 2017;27(5):2739–2757. doi: 10.1093/cercor/bhw113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kravitz DJ, Saleem KS, Baker CI, Ungerleider LG, Mishkin M. The ventral visual pathway: An expanded neural framework for the processing of object quality. Trends in Cognitive Sciences. 2013;17(1):26–49. doi: 10.1016/j.tics.2012.10.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lerner TN, Holloway AL, Seiler JL. Current Opinion in Neurobiology. Vol. 67. Elsevier Ltd; 2021. Dopamine, Updated: Reward Prediction Error and Beyond; pp. 123–130. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Li Z, Yan A, Guo K, Li W. Fear-related signals in the primary visual cortex. Current Biology. 2019;29 doi: 10.1016/j.cub.2019.09.063. [DOI] [PubMed] [Google Scholar]
  32. Liu C, Goel P, Kaeser PS. Spatial and temporal scales of dopamine transmission. Nature Reviews Neuroscience. 2021;22(6):345–358. doi: 10.1038/s41583-021-00455-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Margolis EB, Lock H, Hjelmstad GO, Fields HL. The ventral tegmental area revisited: Is there an electrophysiological marker for dopaminergic neurons? Journal of Physiology. 2006;577(3):907–924. doi: 10.1113/jphysiol.2006.117069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Murris SR, Arsenault JT, Raman R, Vogels R, Vanduffel W. Electrical stimulation of the macaque ventral tegmental area drives category-selective learning without attention. Neuron. 2021;109(8):1381–1395.:e7. doi: 10.1016/j.neuron.2021.02.013. [DOI] [PubMed] [Google Scholar]
  35. Pignatelli M, Bonci A. Role of Dopamine Neurons in Reward and Aversion: A Synaptic Plasticity Perspective. Neuron. 2015;86(5):1145–1157. doi: 10.1016/j.neuron.2015.04.015. [DOI] [PubMed] [Google Scholar]
  36. Sakai K, Miyashita Y. Neural organization for the long-term memory of paired associates. Nature. 1991 November;354:152–155. doi: 10.1038/354152a0. [DOI] [PubMed] [Google Scholar]
  37. Salamone JD. Functions of mesolimbic dopamine: Changing concepts and shifting paradigms. Psychopharmacology. 2007;191(3):389. doi: 10.1007/s00213-006-0623-9. [DOI] [PubMed] [Google Scholar]
  38. Salamone JD, Correa M. The Mysterious Motivational Functions of Mesolimbic Dopamine. Neuron. 2012;76(3):470–485. doi: 10.1016/j.neuron.2012.10.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Salamone JD, Correa M, Farrar A, Mingote SM. Effort-related functions of nucleus accumbens dopamine and associated forebrain circuits. Psychopharmacology. 2007;191(3):461–482. doi: 10.1007/s00213-006-0668-9. [DOI] [PubMed] [Google Scholar]
  40. Schultz W. Recent advances in understanding the role of phasic dopamine activity. F1000Research. 2019;8 doi: 10.12688/f1000research.19793.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Schultz W, Stauffer WR, Lak A. The phasic dopamine signal maturing: from reward via behavioural activation to formal economic utility. Current Opinion in Neurobiology. 2017;43:139–148. doi: 10.1016/j.conb.2017.03.013. [DOI] [PubMed] [Google Scholar]
  42. Seitz AR, Kim D, Watanabe T. Rewards evoke learning of unconsciously processed visual stimuli in adult humans. Neuron. 2009;61(5):700–707. doi: 10.1016/j.neuron.2009.01.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Seitz AR, Watanabe T. A unified model for perceptual learning. Trends in Cognitive Sciences. 2005;9(7):329–334. doi: 10.1016/j.tics.2005.05.010. [DOI] [PubMed] [Google Scholar]
  44. Speranza L, di Porzio U, Viggiano D, de Donato A, Volpicelli F. Dopamine: The neuromodulator of long-term synaptic plasticity, reward and movement control. Cells. 2021;10(4) doi: 10.3390/cells10040735. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tamaki M, Berard Av, Barnes-Diana T, Siegel J, Watanabe T, Sasaki Y. Reward does not facilitate visual perceptual learning until sleep occurs. Proceedings of the National Academy of Sciences of the United States of America. 2020;117(2):959–968. doi: 10.1073/pnas.1913079117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Tsushima Y, Seitz A, Watanabe T. Task-irrelevant learning occurs only when the irrelevant feature is weak. Current Biology. 2008;18(12):516–517. doi: 10.1007/s00426-020-01337-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Ungerleider LG, Doyon J, Karni A. Imaging brain plasticity during motor skill learning. Neurobiology of Learning and Memory. 2002;78(3):553–564. doi: 10.1006/nlme.2002.4091. [DOI] [PubMed] [Google Scholar]
  48. Vancraeyenest P, Arsenault JT, Li X, Zhu Q, Kobayashi K, Isa K, Isa T, Vanduffel W. Selective Mesoaccumbal Pathway Inactivation Affects Motivation but Not Reinforcement-Based Learning in Macaques. Neuron. 2020;108(3):568–581.:e6. doi: 10.1016/j.neuron.2020.07.013. [DOI] [PubMed] [Google Scholar]
  49. Varazzani C, San-Galli A, Gilardeau S, Bouret S. Noradrenaline and dopamine neurons in the reward/effort trade-off: A direct electrophysiological comparison in behaving monkeys. Journal of Neuroscience. 2015;35(20):7866–7877. doi: 10.1523/JNEUROSCI.0454-15.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Vogels R. Sources of adaptation of inferior temporal cortical responses. Cortex. 2016;80:185–195. doi: 10.1016/j.cortex.2015.08.024. [DOI] [PubMed] [Google Scholar]
  51. Watanabe T, Sasaki Y. Perceptual learning: toward a comprehensive theory. Annual Review of Psychology. 2015;55(1):197–221. doi: 10.1146/annurev-psych-010814-015214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Xu TX, Yao WD. D1 and D2 dopamine receptors in separate circuits cooperate to drive associative long-term potentiation in the prefrontal cortex. Proceedings of the National Academy of Sciences of the United States of America. 2010;107(37):16366–16371. doi: 10.1073/pnas.1004108107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Zubair M, Murris S, Kaoru I, Hirtaka O, Yoshinori K, Kenta K, Vanduffel W, Tadashi I. Divergent whole brain projections from the ventral midbrain in macaques. Cerebral Cortex. 2021;31(6):2913–2931. doi: 10.1093/cercor/bhaa399. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material
EMS158944-supplement-Supplementary_Material.pdf (617.9KB, pdf)

Data Availability Statement

The behavioral and electrophysiological datasets are available at Mendeley Data: https://data.mendeley.com/v1/datasets/6xkbhhwbxn/draft?preview=1. Custom scripts to generate the figures and statistical outcomes in this study are available in the same online repository. Any additional information and/or raw data can be requested from the senior author upon request.

RESOURCES