Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Mar 26.
Published in final edited form as: Vision Res. 2010 Aug 21;50(24):2683–2691. doi: 10.1016/j.visres.2010.08.020

Effects of saccades on visual processing in primate MSTd

Shaun L Cloherty a, Michael J Mustari b, Marcello GP Rosa c, Michael R Ibbotson a,b,*
PMCID: PMC4374656  NIHMSID: NIHMS673624  PMID: 20732345

Abstract

In surveying their visual environment, primates, including humans make frequent rapid eye movements known as saccades. Saccades result in rapid motion of the retinal image and yet this motion is not perceived. We recorded saccade-related changes in neural activity in the dorsal medial superior temporal area (MSTd) of alert macaque monkeys. We show that the spontaneous activity of neurons in MSTd is modulated around the time of saccades. Some cells show considerable suppression of spontaneous activity, while most show early and significant enhancement. While this modulation of spontaneous activity is variable, the concomitant modulation of neural responses evoked by flashed visual stimuli is uniform and stereotypical – visual responses are suppressed for stimuli presented around the time of saccades and enhanced for stimuli presented afterwards. The combined modulation of spontaneous activity and evoked visual responses likely serves to reduce the detectability of peri-saccadic stimuli and promote the perceptual awareness of visual stimuli between saccades.

Keywords: Visual system, Eye movements, Saccadic suppression, Post-saccadic enhancement, Perception, Macaque cortex

1. Introduction

Evolution has crafted a visual system in primates in which high spatial resolution is only achieved in a small patch of central retina known as the fovea. To see the whole visual field in high resolution the system has evolved simultaneously a highly mobile eye and associated neural circuits to control eye movements (Carpenter, 1988). The eyes need to look to see, and to do so the eyes make saccades (Findlay & Gilchrist, 2003). These pre-planned eye movements shift the eyes very rapidly between visual targets, after which the eye pauses (fixates) to extract the useful information from the static image. The inter-saccadic period usually lasts around 200–400 ms.

How does the visual system cope with the frequent transient shifts of the visual scene associated with saccades? It is common experience that we do not perceive rapid image motion each time we make a saccade, and a reduction in visual sensitivity has been shown quantitatively in many studies (e.g., Burr, Holt, Johnstone, & Ross, 1982; Burr, Morrone, & Ross, 1994; Campbell & Wurtz, 1968; Diamond, Ross, & Morrone, 2000; Holt, 1903; Matin, 1974; Volkmann, Riggs, White, & Moore, 1978). There are two main opposing views to explain how the visual disturbances during saccades are hidden from everyday perception. The first suggests that visual motion processing is actively suppressed during saccades by an extra-retinal mechanism associated with the planning required before making a saccade (for review see Ross, Morrone, Goldberg, & Burr, 2001). The second theory relies on backward temporal masking to explain the phenomenon. Temporal masking refers to the interaction between sequentially presented visual stimuli. In effect, the presence of a post-saccadic stimulus overpowers the perception of the brief intra-saccadic disturbance (Campbell & Wurtz, 1968; Castet, Jeanjean, & Masson, 2002; Ibbotson & Cloherty, 2009). This temporal masking has been termed “saccadic omission” because no actual suppression is required (Campbell & Wurtz, 1968).

Investigations of the visual pathways in non-human primates have shown a consistent trend: neural responses to visual stimuli presented around the time of saccades follow a biphasic pattern (dorsal lateral geniculate nucleus: Reppas, Usrey, & Reid, 2002; parietal cortex: Bremmer, Kubischik, Hoffmann, & Krekelberg, 2009; Ibbotson, Crowder, Cloherty, Price, & Mustari, 2008). Starting around 100 ms before saccade onset, neural responses are suppressed. This suppression is maximal for responses to stimuli presented at saccade onset and persists until saccade end. Furthermore, there is a significant post-saccadic (or re-fixation) enhancement of spiking responses peaking 50–150 ms after saccade end and lasting for 300–400 ms (Bremmer et al., 2009; Ibbotson, Price, Crowder, Ono, & Mustari, 2007; Ibbotson et al., 2008; Rajkai et al., 2008). This post-saccadic enhancement is accompanied by a reduction in response latency (Price, Ibbotson, Ono, & Mustari, 2005; Ibbotson, Crowder, & Price, 2006; Ibbotson et al., 2008). It has been suggested that the pre-saccadic suppression assists in reducing the visibility of the scene during the saccade but does not fully suppress it. The post-saccadic enhancement then acts to promote the saliency of any stimulus present at re-fixation, thus generating robust backward masking of the visual input during a saccade (Ibbotson & Cloherty, 2009). This view, in some respects, unifies the saccadic suppression and backward masking theories.

Here we extend recent studies by considering perceptual omission as a problem of signal detection at the neural level. The premise of the present inquiry is that the ongoing spontaneous activity of visual neurons represents a noisy background against which the response to visual stimuli is to be detected. In this context, modulation of both visual responses and the spontaneous background affect detectability and are therefore likely to determine perceptual performance around the time of saccades. We quantitatively assess the effect of saccades on neural responses to visual stimulation and on the background spontaneous activity of visual neurons in MSTd of macaque parietal cortex, which is a motion-sensitive area that responds optimally to large-field stimuli (Duffy & Wurtz, 1991). MSTd is known to exhibit strong biphasic modulation of neural activity around the time of saccades (Ibbotson et al., 2007, 2008; Bremmer et al., 2009). We show that spontaneous activity of neurons in MSTd is also modulated at the time of saccades. Moreover, many cells exhibit modulation of spontaneous activity seemingly opposite to that of visual responses previously described. We assess the variability in these effects and suggest an interpretation whereby the modulation of visual responses and the prevailing spontaneous activity combine to produce a unified and robust perceptual outcome.

2. Materials and methods

2.1. Surgical procedures

Data were collected from two rhesus macaque monkeys (Macaca mulatta). All surgical and experimental procedures were performed in compliance with NIH guidelines and protocols approved by the Institutional Animal Care and Use Committee at Emory University. Animals were anaesthetised by inhalation of gaseous isoflurane (1.25–2.0%) and, under aseptic conditions, were fitted with an MRI-compatible head stabilization system and recording chamber (Crist Instruments, MD). Scleral search coils for measuring eye movements were implanted beneath the conjunctiva in both eyes. Recording chambers were positioned to provide access to the superior temporal sulcus (Lateral 15 mm; Posterior 5 mm).

Recording site locations were confirmed using magnetic resonance imaging (MRI, Siemens 3-Tesla). During imaging the animals were sedated (ketamine/telezol) and anaesthetised (isoflurane, 1.0–1.5%). Vital signs including blood pressure, heart rate, body temperature, expired CO2 and blood oxygenation were monitored and maintained at physiological levels. Animals were stabilised in an MRI compatible stereotaxic frame (Crist Instruments, MD) and scans were made at 1 mm intervals throughout the anterior–posterior extent of the brain. NeuroLens software (http://www.neuro-lens.org/) was used to identify regions of interest below the recording chambers.

Recording electrodes were inserted via an adjustable radius-and-angle positioning device that attached to the recording chamber. This device included a centring system that carried a saline-filled guide tube made of fused silica (Plastics One Inc., Roanoke VA) for visualization during MRI sessions. Recording tracks vertically penetrated the anterior bank of the superior temporal sulcus. All recording sites were located in the dorsal medial superior temporal area (MSTd).

2.2. Visual stimuli and task

During recording sessions monkeys sat in a primate chair with their heads stabilized in the horizontal stereotaxic plane. Visual stimuli were projected onto a tangent screen placed 61 cm in front of the animals subtending a visual field of 77° × 77°. Stimuli were projected using a Mirage 2000 Digital Light Projector (DLP – Christie Digital, CA) with resolution 1024 × 1024 pixels (frame rate 96 Hz; mean luminance 170 cd/m2). Because the mirrors used to make images in the projector can be moved to an ON or OFF position thousands of times per frame and each pixel in a frame is updated simultaneously, there are no problems with phosphor luminance decay or scanning and refresh flicker as seen with CRT and LCD displays (for discussion of this point, see Ibbotson et al., 2008).

Monkeys were trained to make saccades back-and-forth between two red fixation targets separated by 10° (5° either side of the centre of the screen), alternately presented on an isoluminant mean grey screen. Monkeys made saccades roughly every 2 s and received a fruit juice reward upon fixating either target and then every 0.5–1 s thereafter for maintaining fixation.

For the majority of the time the only visual stimulus present was one of the red fixation targets. A wide field random texture pattern consisting of 0.8° black and white squares was briefly presented for a single frame at random times relative to the saccades (inter-stimulus intervals ranged from 100 to 500 ms). Monkeys were not required to indicate that they had seen the flashed stimuli and were not rewarded for any specific behaviour related to the stimulus. The experimental protocol aimed to provide sufficient saccades to obtain a minimum of six flashes in each 20 ms bin between −200 and +200 ms relative to saccade onset. The monkeys fixated targets for extended periods between saccades. During these periods we recorded control responses (without saccades).

Prior to the saccade experiment we determined the extent of the receptive field, the preferred direction and the preferred speed of each cell using moving sinusoidal gratings and random dot stimuli while the monkey fixated a central spot. Care was taken to ensure the stimulus screen covered the entire receptive field of all recorded neurons at all times. However, the receptive fields of the recorded neurons were sufficiently large as to include the fixation targets. We therefore also tested each cell for spiking activity evoked by the fixation target present during saccades. None of the cells included in this study exhibited spiking responses induced by the fixation targets.

2.3. Data collection

A magnetic search coil system (CNC Electronics, Seattle, WA) was used to measure eye positions in two dimensions (sample rate 1 kHz). Iron-tipped, epoxy-coated tungsten electrodes (Frederick-Haer Corporation, Brunswick, ME) with impedances ranging from 1 to 4 MX were used to record from cells. Neural responses were sampled at 25 kHz and stored for off-line analysis. All signals were digitised with 16-bit precision using a Power 1401 acquisition system (CED, Cambridge, England, UK).

Eye velocity was calculated by differentiation of the eye position signals using a finite difference formula (temporal resolution = 1 ms). Saccade onset was defined as the moment when the eye velocity first exceeded 10°/s. Stimulus onset was determined using a frame synchronous event marker generated by the stimulus computer. Spike arrival times were determined off-line using action potential template matching (Spike2, CED, Cambridge, England, UK).

2.4. Data analysis

Spiking responses are presented as spike density functions (SDFs) with 1 kHz resolution obtained by convolving a Gaussian kernel of unit area and σ = 5 ms with a train of Dirac delta functions, each of which corresponded to a spike. Mean SDFs were calculated by trial averaging responses to individual stimulus presentations.

To quantify the effect of saccades on visual responses in MSTd we first characterised the neural response of each cell to flashes delivered more than 500 ms before or after a saccade (the control condition). For each cell all control trials were aligned at the time of flash onset. An example for one MSTd neuron in our population is shown in Fig. 1A. The onset and duration (10 ms) of the flashed stimulus is indicated by the solid black bar. The spontaneous activity for each cell in the absence of saccades was estimated from the mean spike rate in the first 25 ms after presentation of the flash as shown by the shaded region in Fig. 1A.

Fig. 1.

Fig. 1

Open in a new tab

Methods. (A) Neural response of a single neuron in MSTd following briefly presented stimuli delivered more than 500 ms before or after a saccade (the control condition). The upper panel shows the raw spike arrival times as a raster plot (n = 938). The lower panel shows the mean spike rate as a spike density function (see Methods). All trials are aligned at the time of stimulus onset. The onset and duration of the flashed stimulus is indicated by the solid black bar. The spontaneous firing rate (dashed line) in the absence of saccades was estimated from the mean spike rate in the first 25 ms after presentation of the stimulus as shown by the shaded region. Visual response amplitude in the absence of saccades was defined as the increase in spike rate above the estimated spontaneous rate. (B) Neural response from the same cell to the same flashed stimuli delivered 20–40 ms after saccade onset, i.e., within a 20 ms time bin centred 30 ms after saccade onset. The upper panel shows the eye position traces. The lower panel shows the mean spike rate as a spike density function (n = 12). All trials are aligned with respect to stimulus onset and the mean response then aligned 30 ms after saccade onset. The onset and duration of the flashed stimulus is indicated by the solid black bar. The prevailing spontaneous rate within each 20 ms time bin (defined relative to saccade onset) was estimated from the mean spike rate in the first 25 ms after presentation of the flashes within that bin. Visual response amplitude was calculated as the amplitude of the evoked response minus the prevailing spontaneous rate within the time bin containing the response. In the example shown, this spontaneous rate (indicated by the dash line) was estimated from the 25 ms period after flashes presented (during other trials) within the shaded region.

The onset of the cell's visual response was calculated by Poisson analysis of the mean spike rate. We fitted a Poisson distribution to the mean spike rate in the spontaneous activity window (shaded region in Fig. 1A). We then calculated a threshold for response onset as the 99% cut-off of this Poisson distribution. Response onset was then defined as the first point at which the mean spike rate exceeded this threshold for a period of at least 25 ms. Visual response amplitude in the absence of saccades was defined as the peak amplitude of the evoked response (after response onset) minus the estimated spontaneous rate. Response latency of the cell was defined as the interval between stimulus onset and the peak of the response. The visual response amplitude for the example cell is indicated in Fig. 1A. The estimated spontaneous rate is indicated by the dashed line.

To assess the effect of saccades, we characterised neural responses to the same flashed stimulus delivered before, during and after saccades. Trials were combined within 20 ms bins based on the time of stimulus presentation relative to saccade onset. Fig. 1B shows eye position (upper trace) and the mean spiking response from the same example neuron for flashes presented 30 ms after a saccade. As for the control condition, the prevailing spontaneous rate for each time bin was estimated from the mean spike rate in the first 25 ms after presentation of the flash. This spontaneous rate was assigned to the 20 ms time bin in which the stimuli were presented (defined relative to saccade onset).

Visual responses were assigned to the time bin corresponding to the time of stimulus presentation plus the response latency of the cell. The example cell has a response latency of 57 ms (Fig. 1A). The visual response for flashes presented 30 ms after a saccade (Fig. 1B) was therefore assigned to the +90 ms time bin (30 + 57 = 87 ms). The amplitude of the visual response was calculate as the peak amplitude of the evoked response minus the prevailing spontaneous rate at the time of the response. In the example shown, this is the spontaneous activity estimated from the 25 ms immediately following flashes presented in the +90 ms bin (shown shaded in Fig. 1B). In this way, we were able to separate changes in the prevailing spontaneous firing rate from changes in the evoked response amplitude.

To quantify the effect of saccades on visual responses, we normalized the visual response amplitude in each bin by dividing by the visual response amplitude in the control condition (in the absence of saccades). Similarly, to quantify the effect of saccades on the spontaneous background, we normalized the spontaneous rate for each bin by dividing by the spontaneous rate in the control condition. Throughout the paper normalized responses are presented on log scale to appropriately represent suppression as strongly as enhancement. In some time bins exhibiting substantial suppression of spiking activity, subtraction of the prevailing spontaneous rate from the evoked visual response led to values close to or less than zero. Data points for these time bins are plotted in grey, outside the limits of the axes.

3. Results

3.1. Modulation of responses to flashed visual stimuli

Normalized visual response amplitudes for three example cells are shown as functions of time relative to saccade onset in Fig. 2. Here, time relative to saccade onset was divided into 20 ms bins as described above. The responses to all flashes presented within each bin were aligned with respect to flash onset and averaged together. Example responses from three different time bins, represented as spike density functions, are shown inset above Fig. 2A. During fixation, this cell responded to the flashed stimuli with a robust burst of spikes approximately 100 ms after presentation of the flash. Consider now the left most response shown inset in Fig. 2A. Here, flashed stimuli were presented 170 ms prior to saccade onset. However, the visual response was observed approximately 100 ms after the flash, i.e., 70 ms prior to saccade onset. This response is indistinguishable from that observed in the control condition. It is therefore represented by the value near unity, plotted at −70 ms (relative to saccade onset at 0 ms).

Fig. 2.

Fig. 2

Open in a new tab

Modulation of visual response amplitude around the time of saccades. (A–C). Normalized visual response amplitude of three neurons as functions of time relative to saccade onset. The visual response amplitude in each 20 ms bin was calculated by subtracting the prevailing spontaneous rate from the peak amplitude of the mean response within that bin. Visual response amplitude was then normalized to the visual response amplitude observed in the control condition (horizontal line). In (A) the cell shows suppression of responses arriving in MSTd at saccade onset and strong post-saccadic enhancement of responses arriving 100–400 ms after saccade onset. Example spike density functions from three different time bins are shown inset for flashes that occurred at −170 ms, −110 ms and +130 ms relative to saccade onset. (B) An example cell showing little in the way of early suppression but exhibiting substantial post-saccadic enhancement. (C) A cell showing strong suppression for flashes presented before saccades but no subsequent post-saccadic enhancement. The grey shaded region shows the mean duration of the saccades. Error bars indicate the standard error of the estimate of the mean response amplitudes. In some time bins, particularly those exhibiting substantial suppression, subtraction of the spontaneous rate resulted in response amplitudes close to or even less than zero. These cannot be represented on the log scale and so are shown in grey below the axes. These points show times at which extreme suppression occurred.

It is evident in Fig. 2 that responses to flashes presented long before saccades are not significantly different from control values (i.e., close to unity). The first cell (Fig. 2A) shows almost complete suppression of the response amplitude for responses arriving in MSTd at saccade onset. This suppression persists until 110 ms after saccade onset. As with the majority of cells, this neuron also showed strong post-saccadic enhancement (100–400 ms after saccade onset). This suppression and subsequent enhancement is evident in the example responses shown inset in Fig. 2A. The second cell (Fig. 2B) shows little in the way of early suppression but exhibits substantial post-saccadic enhancement. The responses of a relatively unusual cell are shown in Fig. 2C. This cell exhibits strong suppression for flashes presented before saccades – manifest as smaller-than-control visual response amplitudes observed 50–70 ms after saccade onset. However, this cell is unusual in that it exhibits no post-saccadic enhancement.

Fig. 3 shows normalized visual response amplitude as a function of time relative to saccade onset, together with example responses for another example cell. This cell exhibits only weak responses to visual stimulation except in the wake of a saccade. This cell's response to flashes presented long before saccades is shown inset (left) in Fig. 3. This response is statistically indistinguishable from that observed during fixation and fails to exceed the 99% Poisson threshold for response onset (indicated by the dashed line in the insets). Under these conditions, this cell would be classed as non-visual. However, after a saccade the cell was highly visual, producing robust spiking responses to the flashed stimuli as shown inset (right) in Fig. 3. Four cells of this type were isolated in addition to the 72 cells presented in the population data that follows.

Fig. 3.

Fig. 3

Open in a new tab

Maximum influence of post-saccadic enhancement. Normalized visual response amplitude as a function of time relative to saccade onset. This cell exhibits only weak responses to visual stimulation during fixation and robust responses after saccades. (Inset, top left) Evidence for a visual response is weak, with the mean response failing to reach the 99% Poisson threshold (dashed line) used to determine response onset (n = 772). (Inset, top right) Mean spike density function showing the response to flashed stimuli presented 120 ms after saccade onset: the visual response is robust (n = 16). The grey shaded region shows the mean duration of the saccades. Time bins exhibiting extreme suppression, resulting in response amplitudes close to or even less than zero, are shown as grey symbols outside the limits of the axes. Error bars indicate the standard error of the estimate of the mean response amplitudes.

While showing examples from individual cells is a useful approach, it does not allow a comprehensive survey of variability within the cell population. We established that on average, across the population, cells exhibit suppression within an early window from 0 to 100 ms after saccade onset. In contrast, on average, enhancement was confined to a late window from 100 to 250 ms after saccade onset. We calculated the mean normalized visual response amplitude within the early (0–100 ms) and late (100– 250 ms) windows for each cell. To illustrate the level of variability within the cell population, we plot these two metrics (early vs. late) for 72 cells in Fig. 4. It is evident that there are no clear subpopulations of cells. Rather, most cells fall in a cluster in the bottom right of the scatter plot. Approximately half the cells in our population (39 cells; 54%) lay in the lower right quadrant. These cells show the classic biphasic pattern of modulation (Ibbotson et al., 2008): i.e., suppression in the early window and prolonged enhancement in the late window. Overall, 49 cells (68%) showed post-saccadic enhancement in the late window (i.e., all cells in the right half-plane). Fifty-seven cells (79%) showed significant suppression in the early window (i.e., all cells in the lower quadrants). For the population as a whole, the resulting effect is a biphasic pattern of modulation with early saccadic suppression followed by post-saccadic enhancement. A relatively small number of cells showed only early suppression (lower left quadrant) or only post-saccadic enhancement (upper right quadrant).

Fig. 4.

Fig. 4

Open in a new tab

Saccade-related modulation of visual responses for 72 neurons in MSTd. The scatter plot shows the mean normalized visual response amplitude in the early response window (0–100 ms after saccade onset) plotted against that in the late response window (100–250 ms) for 72 cells. The majority of cells (n = 39) are clustered in the lower right quadrant of the scatter plot, indicative of early suppression followed by late enhancement of visual responses. Cells for which the mean response amplitude was very close to or even less than zero, thus preventing plotting on a log scale, are shown as grey symbols outside the limits of the axes. These points show cells exhibiting extreme suppression of visual responses.

3.2. Modulation of spontaneous activity

The data presented above illustrate a consistent pattern of modulation during saccades for visual responses in MSTd. However, modulation of visual responses is not the only factor driving visual performance. Modulation of the spontaneous background spiking activity likewise limits the capacity of the visual system to detect visual signals.

To quantify this effect we measured the ongoing spontaneous activity of neurons around the time of saccades as outlined above. Specifically, spontaneous firing rate was estimated for each 20 ms bin and normalized by dividing by the mean spontaneous activity of each cell in the absence of saccades. Fig. 5 shows normalized spontaneous rate as a function of time relative to saccade onset for three example cells. The cells shown in Fig. 5A and B show substantial enhancement of spontaneous rate in the early window from 0 to 100 ms after saccade onset with less substantial enhancement in the late window. The third cell, shown in Fig. 5C again shows substantial enhancement in the early window but also shows considerable suppression of spontaneous firing throughout the late window. Early enhancement of spontaneous firing, as revealed by these cells, was by far the most common effect in our cell population. However, roughly a third of cells showed little in the way of enhancement, and rather, exhibit suppression of spontaneous firing in the early window.

Fig. 5.

Fig. 5

Open in a new tab

Modulation of spontaneous activity around the time of saccades. (A–D) Normalized spontaneous firing rate of three neurons as functions of time relative to saccade onset. Spontaneous firing rate was estimated within each 20 ms bin and normalized by dividing by the mean spontaneous rate of each cell in the absence of saccades (as represented by the horizontal line). The cells in (A) and (B) show substantial enhancement of spontaneous rate from 0 to 100 ms after saccade onset. (C) An example cell showing substantial enhancement in the early window (0–100 ms) followed by suppression of spontaneous firing until approximately 300 ms after saccade onset. Early enhancement of spontaneous firing as shown by these cells was observed in more than half the cells (47 cells, 65%). Error bars indicate the standard error of the estimate of the mean response amplitudes. The grey shaded region shows the mean duration of the saccades. Grey symbols indicate periods of extreme suppression (see Fig. 2; caption).

To illustrate the diversity of effects across our cell population, we calculated the mean spontaneous rate in both the early (0–100 ms) and late (100–250 ms) windows as described above. These two metrics for each cell in our population are plotted (early vs. late) in Fig. 6. It is immediately apparent that the points are far more scattered here than was the case for the visual response amplitudes (Fig. 4). This reflects the increased diversity of effects seen in the modulation of spontaneous activity around the time of saccades. Surprisingly, the distribution of points appears in some respects opposite to that shown for the visual responses in Fig. 4. Forty-seven cells (65%; upper left and right quadrants) showed a significant increase in spontaneous firing in the early window. This is in contrast to the modulation of visual responses where the majority of cells showed a substantial reduction in the early window (Fig. 4). Almost half the cells (35 cells; 49%; upper and lower left quadrants) showed a significant reduction in spontaneous firing throughout the late window where, for the majority of cells, visual responses were enhanced (Fig. 4).

Fig. 6.

Fig. 6

Open in a new tab

Saccade-related modulation of spontaneous activity for 72 neurons in MSTd. The scatter plot shows the mean normalized spontaneous rate in the early window (0–100 ms after saccade onset) plotted against that in the late response window (100–250 ms). In contrast to the modulation of visual responses shown in Fig. 4, it is evident from the lack of clustering that cells in MSTd exhibit considerable variability in saccade-related modulation of spontaneous activity. Cells for which the mean spontaneous rate was very low or even completely suppressed in either window are shown as grey symbols outside the limits of the axes. These points show cells exhibiting extreme suppression of spontaneous activity.

3.3. Reconciling opposing modulatory effects

From the data presented above it is evident that visual signals observed in MSTd are substantially modified around the time of saccades. It is also clear that the ongoing spontaneous activity observed in MSTd is also modulated at the same time. However it is unclear how these two effects interact or how they might determine performance of the visual system during saccades. In Fig. 7A we present normalized total response amplitude as a function of time relative to saccade onset, averaged across 72 cells. The total response amplitude reflects the peak spike rate elicited by the flashed stimuli, without subtraction of the prevailing spontaneous rate. As in Figs. 2–4, total response amplitudes are plotted at the time the responses were observed in MSTd. It is evident that, on average, the peak spike rate is reduced for responses observed in MSTd immediately after saccades. This period of suppression is then followed by a prolonged period during which the peak firing rate in MSTd is enhanced.

Fig. 7.

Fig. 7

Open in a new tab

Population averages. (A) Normalized total response amplitude as a function of time relative to saccade onset, averaged across 72 cells. The total response is measured as the peak spike rate, including the evoked visual response and the ongoing spontaneous activity. On average, peak spike rate exhibits a biphasic modulation: responses arriving in MSTd immediately after saccades are suppressed while later responses are enhanced. (B) Normalized spontaneous activity over the same time period, averaged across 72 cells. It is evident that on average cells in MSTd exhibit a significant increase in spontaneous activity immediately after saccades. (C) Normalized visual response amplitude (i.e., total response amplitude minus the prevailing spontaneous rate) normalized and averaged across 72 cells. On average, neurons in MSTd exhibit significant suppression followed by enhancement of visual response amplitudes beginning immediately after saccades. Note that the level of suppression is far greater (more than 80% compared to the control condition) than that seen for the total response (30–50% compared to the control condition). This is because spontaneous activity in MSTd is enhanced immediately after saccades, precisely when the suppressed visual signals are arriving in MSTd. In each panel, the level of significance of the modulation in each 20 ms bin is indicated by stars (t-tests; *, p < 0.05; **, p < 0.01). Error bars indicate the standard error of the estimate of the mean. The grey shaded region shows the mean duration of the saccades.

For comparison, in Fig. 7B we show the normalized spontaneous rate within each 20 ms bin, again averaged across the same 72 cells. The average spontaneous activity profile reveals a number of interesting features. Firstly, on average, the population shows no significant suppression of spontaneous activity, even though suppression was observed in many individual cells (Fig. 6). Secondly, on average, the population shows a significant increase in spontaneous activity immediately after saccades. That is, during the same period in which the peak firing rate in response to visual stimuli appeared to be significantly reduced (Fig. 7A). To separate changes in spontaneous firing from changes in the visual response evoked by the flashed stimuli, in Fig. 7C we show the mean normalized visual response amplitudes observed in MSTd, i.e., total response amplitude minus the prevailing spontaneous rate, normalized and averaged across all cells in the population. Here we see that, like peak firing rate, the visual component is significantly suppressed initially. On average visual responses were suppressed by more than 80% compared to the control condition (Fig. 7C). This compares with suppression of only 30–50% of the total response amplitude, i.e., when the increase in spontaneous activity is included (Fig. 7A).

It is interesting that spontaneous activity is enhanced in the early window (0–100 ms) after saccade onset, while in this same window, visual responses observed in MSTd are suppressed. Additional insight may be gained by considering the modulatory effects outlined above in terms of their effect on the detectability of the evoked visual response – specifically, the resulting evoked-to-spontaneous ratio. The implicit assumption here is that the ongoing spontaneous activity represents a noisy background against which the visual response is to be detected, and that it is the evoked-to-spontaneous ratio which limits detection. Fig. 8A shows the evoked-to-spontaneous ratio in each 20 ms bin, averaged across all cells in our population. It is evident that on average, the evoked-to-spontaneous ratio is transiently reduced immediately following a saccade and is subsequently enhanced.

Fig 8.

Fig 8

Open in a new tab

Detectability of visual responses in MSTd around the time of saccades. (A) Variation in the evoked-to-spontaneous ratio as a function of time relative to saccade onset, averaged across 72 cells. The evoked-to-spontaneous ratio, an indicator of detectability of the evoked visual response, is transiently reduced immediately following a saccade and is then subsequently enhanced. (B) Variation in the evoked-to-spontaneous ratio over the same time period averaged only across those cells which exhibited suppression of their spontaneous activity 100–250 ms after saccade onset (35 cells, 49% of the population). It is evident that this suppression of spontaneous activity produces a substantial increase in the evoked-to-spontaneous ratio in this window. Notably, these cells also show a substantial reduction in the evoked-to-spontaneous ratio during and immediately following saccades. (C) Variation in the evoked-to-spontaneous ratio averaged only across those cells which exhibited an enhancement of their spontaneous activity 100–250 ms after saccade onset (37 cells, 51% of the population). These cells also exhibit a transient reduction in the evoked-to-spontaneous ratio immediately after saccades. However, despite the subsequent increase in spontaneous rate these cells also show a substantial post-saccadic increase in their evoked-to-spontaneous ratio. In each panel, the level of significance of the modulation in each 20 ms bin is indicated by stars (t-tests; *, p < 0.05; **, p < 0.01). The grey shaded region shows the mean duration of the saccades.

Changes in the evoked-to-spontaneous ratio reflect changes in both the visual response amplitude and in the prevailing spontaneous rate. As noted previously, approximately half of cells in our population show suppression of spontaneous activity during the late window from 100 to 250 ms after saccade onset while the remainder shows some level of enhancement. It is therefore plausible that the post-saccadic enhancement in the evoked-to-spontaneous ratios seen in Fig. 8A simply reflects the reduction in spontaneous activity seen in half our population rather than a more general effect across the whole population. To explore this possibility, we partitioned our population into those cells that showed suppression of spontaneous activity in the late window (cells in the upper and lower left quadrants in Fig. 6) and those that showed enhancement (cells in the upper and lower right quadrants in Fig. 6). Average evoked-to-spontaneous ratios within these groups are shown in Fig. 8B (late suppression) and C (late enhancement), respectively. Cells plotted in Fig. 8B show a substantial reduction in the evoked-to-spontaneous ratio before, during, and immediately following saccades. In the late response window suppression of spontaneous activity produces a substantial increase in the evoked-to-spontaneous ratio (Fig. 8B). For the remainder of cells (Fig. 8C) there is a transient reduction in the evoked-to-spontaneous ratio, but this ‘reduced-detectability’ phase appears later than occurred for the cell group in Fig. 8B. Despite an increase in spontaneous rate in the late response window, these cells show a substantial post-saccadic increase in evoked-to-spontaneous ratios (Fig. 8C).

4. Discussion

4.1. Modulation of spontaneous activity

When monkeys made saccades between fixation targets across a blank screen the ongoing activity in most MSTd neurons was modulated. Across the population, the dominant effect was a transient increase in spontaneous activity during and soon after saccades. Subsequent to this the modulation of spontaneous activity is more variable, with the population divided almost equally between cells showing suppression and those showing enhancement. Ibbotson et al. (2008) measured the modulation of firing activity in MSTd during saccades in total darkness – the monkeys made saccades in response to auditory cues, without any visual input at all. At the population level neurons in MSTd showed early post-saccadic enhancement of spontaneous activity, similar to that shown in Fig. 7B, and no significant suppression. It was recently shown that neurons in four regions of primate parietal cortex, including MSTd, showed a systematic relationship between mean firing rate and the position of the eyes in the orbit (Morris, Kubischik, Hoffmann, Krekelberg, & Bremmer, 2010). This suggests that some of the increase in spiking activity that we report at the time of saccades in the absence of visual stimulation could be related to an eye position signal.

Bremmer et al. (2009) also recorded spontaneous activity in area MSTd during saccades. Based on the population average, they reported a small, short-lived peri-saccadic suppression of spontaneous spiking activity and no enhancement for saccades made in the absence of strong visual input. It is necessary to reconcile these data with those reported above. First, Bremmer and colleagues measured spontaneous activity with the monkeys in very dim light, while spontaneous activity in our experiments was measured in bright light or total darkness (Ibbotson et al., 2008). Baseline spontaneous rates may well be different for dark and light conditions, but as different populations of cells were recorded this comparison is not possible. If baseline levels change it is plausible that modulations during saccades are similarly different. Taking a functional viewpoint, it is possible that the system operates differently in darkness, dim light and bright light. In the latter there is a far greater need to suppress motion of the background scene because it is highly visible. Psychophysical evidence suggests that at low luminance, contrast sensitivity is enhanced rather than suppressed during saccades (Burr et al., 1982). It is reasonable to suggest that differences in visual system performance, driven by dark and light adaptation respectively, also lead to different saccadic suppression strategies. Second, given the bright conditions in the present experiments it might be suggested that we are observing small responses to stimuli outside the classical receptive fields of the MSTd neurons. However, this is highly unlikely because the onset of the increased spiking activity occurred at or before saccade onset. Given that MSTd neurons have latencies of 40–100 ms, any motion related visual stimulus caused by the movement of the eye during the saccade (duration 30 ms) would not lead to measurable changes in spiking activity until well after saccade end. Third, it is important to note that approximately one third of the cells in the present work (25 cells; 35%) do exhibit suppression of their spontaneous activity in the early post-saccadic time window comparable to that reported by Bremmer and colleagues. However, at the population level this suppression was masked by the transient enhancement seen in the other two thirds of cells. Moreover, some suppression was seen in the cell population in which responses were measured in darkness (Ibbotson et al., 2008). It is likely that sampling differences had some influence on the population averages in the three studies. This, combined with the differences in visual environment, make direct comparisons between the studies difficult.

4.2. Modulation of perceptual awareness

The modulation of visual response amplitudes in MSTd is quite strong. Suppression, on average was 80–100% of the control value (i.e., visual responses were reduced to less than 20% of their control amplitudes). Despite this, when the total response (i.e., peak spike rate) was measured, including any modulation in the intrinsic spontaneous activity of MSTd we found that the suppression appeared far smaller (reductions of only 30–50%). It would appear that although saccadic suppression of visual responses is evident throughout the visual system, the modulation of spontaneous activity which we observe in MSTd negates some of this suppression. This suggests a complex interplay between saccadic modulation and visual performance.

The apparent contradictory influences on neurons in MSTd may be reconciled by noting their effect on the evoked-to-spontaneous ratio and hence the detectability of visual signals around the time of saccades. In the early window, from 0 to 100 ms after saccade onset, visual response amplitude (the signal) is suppressed (Fig. 7C). This suppression is substantial, but by no means absolute. During this same period, background spontaneous firing is enhanced (Fig. 7B). Both effects serve to reduce the evoked-to-spontaneous ratio and hence reduce the detectability of the signal compared to that during fixation. In the subsequent period, from 100 to 250 ms after saccade onset, visual response amplitude is enhanced while spontaneous background firing is significantly reduced in approximately half the population. In the late window these effects again combine, this time to increase the evoked-to-spontaneous ratio, in turn improving detectability of the visual signal. The effect of modulation of spontaneous activity on the detectability of visual signals as posited here is analogous to that demonstrated in relay neurons of the LGN (Guido, Lu, Vaughan, Godwin, & Sherman, 1995). There, sustained spontaneous activity was shown to reduce detectability of visual responses while a reduction in spontaneous activity served to improve detectability. Our results suggest that in MSTd, the detectability of visual input during saccades is reduced while detectability of visual input at re-fixation is enhanced. These effects likely subserve the perceptual omission of peri-saccadic visual stimuli (Ibbotson & Cloherty, 2009).

4.3. Direct vs. inherited effects

Is MSTd itself actively modulated during saccades? Alternatively, are these effects inherited from earlier visual areas which are themselves modulated by saccade-related corollary discharges (Sperry, 1950; von Holst, 1954)? The present data reveal that saccadic modulation of visual response amplitude is far more predictable than that of spontaneous activity in the absence of visual input. The great majority of cells exhibit a classic biphasic response modulation of visual responses. The difference between this stereotypical pattern and that seen in the absence of visual input is intriguing. It is plausible that a stereotypical biphasic modulation occurs in earlier visual regions and that this propagates throughout the visual system. In contrast, the modulation of spontaneous activity might suggest that MSTd is directly modulated by a variety of inputs, including those from the preceding visual pathways. The diversity in effects seen in the modulation of spontaneous activity suggest that MSTd could be directly modulated by both suppressive and excitatory signals during saccades (Barone & Kennedy, 2000; Higo, Akashi, Sakimura, & Tamamaki, 2009). Certainly, input arises from some internal mechanism since the modulation is seen during saccades in total darkness (Ibbotson et al., 2008).

There is evidence that neurons in the first brain region of the retino-cortical pathway, the dorsal lateral geniculate nucleus (LGNd), are influenced by corollary discharge (cats: Lee & Malpeli, 1998; monkeys: Ramcharan, Gnadt, & Sherman, 2001; Reppas et al., 2002; Royal, Sary, Schall, & Casagrande, 2006). There is pre-saccadic suppression and post-saccadic enhancement in LGNd, even in darkness (Royal et al., 2006). Psychophysical evidence in humans also suggests very early saccadic and blink suppression, perhaps in LGNd (Bristow, John-Dylan Haynes, Sylvester, & Frith, 2005; Burr et al., 1994; Thilo, Santoro, Walsh, & Blakemore, 2004). However, peri-saccadic visual stimuli although omitted from perceptual awareness can alter the perception of subsequent stimuli (Ibbotson & Cloherty, 2009; Watson & Krekelberg, 2009). This suggests that peri-saccadic visual input is processed by the brain and that suppression is not restricted to the LGN alone (for a mini-review, see Burr, 2005). Chahine and Krekelberg (2009) suggest that the cortex ‘. . . regulates the amount of suppression, while the LGN performs the actual suppression’ and it is likely that this is true for some visual pathways.

Our data are certainly consistent with the notion that visual signals are modulated before arriving in the parietal cortex, but that capacity exists for further modification of these signals in parietal cortex. Wide-spread modulation of brain activity has been reported in the human brain during blinks, which appear to evoke a similar mechanism of suppression to saccades (Bristow et al., 2005). Bremmer et al. (2009) have reported similar biphasic saccade-related modulations to those reported here in the medial temporal (MT) and ventral intraparietal (VIP) areas of parietal cortex, which code the direction and speed of image motion (e.g., Price, Ono, Mustari, and Ibbotson, 2005; Zhang, Heuer, & Britten, 2004). However, in contrast to MSTd ongoing spontaneous activity in these areas appears largely unaffected by saccades. It is plausible that any direct modulation of parietal cortex differs between brain areas.

5. Conclusion

Visual responses in MSTd are modulated around the time of saccades in a stereotypical manner: responses to stimuli presented immediately before or during saccades are suppressed while responses to stimuli presented after saccades are enhanced. However, the ongoing non-visual spontaneous activity in MSTd neurons is modulated in a far less predictable manner. Overall, there appears to be interplay between modulation in the background firing rate (spontaneous activity) and the evoked visual responses in MSTd. This interplay appears to reduce detectability of visual stimuli before and during saccades and to enhance detectability of visual stimuli in the inter-saccade period.

References

  1. Barone P, Kennedy H. Non-uniformity of neocortex: Areal heterogeneity of NADPH-diaphorase reactive neurons in adult macaque monkeys. Cerebral Cortex. 2000;10(2):160–174. doi: 10.1093/cercor/10.2.160. [DOI] [PubMed] [Google Scholar]
  2. Bremmer F, Kubischik M, Hoffmann KP, Krekelberg B. Neural dynamics of saccadic suppression. Journal of Neuroscience. 2009;29:12374–12383. doi: 10.1523/JNEUROSCI.2908-09.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bristow D, John-Dylan Haynes JD, Sylvester R, Frith CD. Blinking suppresses the neural response to unchanging retinal stimulation. Current Biology. 2005;15:1296–1300. doi: 10.1016/j.cub.2005.06.025. [DOI] [PubMed] [Google Scholar]
  4. Burr DC. Vision: In the blink of an eye. Current Biology. 2005;15:R554–R556. doi: 10.1016/j.cub.2005.07.007. [DOI] [PubMed] [Google Scholar]
  5. Burr DC, Holt J, Johnstone JR, Ross J. Selective depression of motion sensitivity during saccades. Journal of Physiology (London) 1982;333:1–15. doi: 10.1113/jphysiol.1982.sp014434. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Burr DC, Morrone MC, Ross J. Selective suppression of the magnocellular visual pathway during saccadic eye-movements. Nature. 1994;371:511–513. doi: 10.1038/371511a0. [DOI] [PubMed] [Google Scholar]
  7. Campbell FW, Wurtz RH. Saccadic omission – why we do not see a gray-out during a saccadic eye-movement. Vision Research. 1968;18:1297–1303. doi: 10.1016/0042-6989(78)90219-5. [DOI] [PubMed] [Google Scholar]
  8. Carpenter RHS. Movements of the eyes. 2nd ed. Pion; London: 1988. [Google Scholar]
  9. Castet E, Jeanjean S, Masson GS. Motion perception of saccade-induced retinal translation. Proceedings of the National Academy of Sciences of the United States of America. 2002;99:15159–15163. doi: 10.1073/pnas.232377199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chahine G, Krekelberg B. Cortical contributions to saccadic suppression. PLoS ONE. 2009;4:e6900. doi: 10.1371/journal.pone.0006900. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Diamond MR, Ross J, Morrone MC. Extraretinal control of saccadic suppression. Journal of Neuroscience. 2000;20:3449–3455. doi: 10.1523/JNEUROSCI.20-09-03449.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Duffy CJ, Wurtz RH. Sensitivity of MST neurons to optic flow stimuli.1. A continuum of response selectivity to large-field stimuli. Journal of Neurophysiology. 1991;65:1329–1345. doi: 10.1152/jn.1991.65.6.1329. [DOI] [PubMed] [Google Scholar]
  13. Findlay JM, Gilchrist ID. Active vision: The psychology of looking and seeing. Oxford University Press; Oxford, UK: 2003. p. 220. [Google Scholar]
  14. Guido W, Lu S-M, Vaughan JW, Godwin DW, Sherman SM. Receiver operating characteristic (ROC) analysis of neurons in the cat’s lateral geniculate nucleus during tonic and burst response mode. Visual Neuroscience. 1995;12(4):723–741. doi: 10.1017/s0952523800008993. [DOI] [PubMed] [Google Scholar]
  15. Higo S, Akashi K, Sakimura K, Tamamaki N. Subtypes of GABAergic neurons project axons in the neocortex. Frontiers Neuroanatomy. 2009;3(25) doi: 10.3389/neuro.05.025.2009. doi:10.3389/neuro.05.025.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. von Holst E. Relations between the central nervous system and the peripheral organs. British Journal of Animal Behaviour. 1954;2:89–94. [Google Scholar]
  17. Holt EB. Eye movements and central anaesthesia. Psychological Review. 1903;4:3–45. [Google Scholar]
  18. Ibbotson MR, Cloherty SL. Visual perception: Saccadic omission – Suppression or temporal masking? Current Biology. 2009;19:R493–496. doi: 10.1016/j.cub.2009.05.010. [DOI] [PubMed] [Google Scholar]
  19. Ibbotson MR, Crowder NA, Price NSC. Neural basis of time changes during saccades. Current Biology. 2006;16:R834–R836. doi: 10.1016/j.cub.2006.08.081. [DOI] [PubMed] [Google Scholar]
  20. Ibbotson MR, Crowder NA, Cloherty SL, Price NSC, Mustari MJ. Saccadic modulation of neural responses: Possible roles in saccadic suppression, enhancement and time compression. Journal of Neuroscience. 2008;28:10952–10960. doi: 10.1523/JNEUROSCI.3950-08.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Ibbotson MR, Price NSC, Crowder NA, Ono S, Mustari MJ. Enhanced motion sensitivity follows saccadic suppression in the superior temporal sulcus of the macaque cortex. Cerebral Cortex. 2007;17:1129–1138. doi: 10.1093/cercor/bhl022. [DOI] [PubMed] [Google Scholar]
  22. Lee D, Malpeli JG. Effects of saccades on the activity of neurons in the cat lateral geniculate nucleus. Journal of Neurophysiology. 1998;79(2):922–936. doi: 10.1152/jn.1998.79.2.922. [DOI] [PubMed] [Google Scholar]
  23. Matin E. Saccadic suppression – review and an analysis. Psychological Bulletin. 1974;81:899–917. doi: 10.1037/h0037368. [DOI] [PubMed] [Google Scholar]
  24. Morris AP, Kubischik M, Hoffmann KP, Krekelberg B, Bremmer F. Dynamics of eye position signals in macaque dorsal areas explain peri-saccadic mislocalization.. Vision Sciences Society Annual Meeting; 2010.2010. [Google Scholar]
  25. Price NSC, Ibbotson MR, Ono S, Mustari MJ. Rapid processing of retinal slip during saccades in macaque area MT. Journal of Neurophysiology. 2005;94:235–246. doi: 10.1152/jn.00041.2005. [DOI] [PubMed] [Google Scholar]
  26. Price NSC, Ono S, Mustari MJ, Ibbotson MR. Comparing acceleration and speed tuning in macaque MT: Physiology and modelling. Journal of Neurophysiology. 2005;94:3451–3464. doi: 10.1152/jn.00564.2005. [DOI] [PubMed] [Google Scholar]
  27. Rajkai C, Lakatos P, Chen CM, Pincze Z, Karmos G, Schroeder CE. Transient cortical excitation at the onset of visual fixation. Cerebral Cortex. 2008;18:200–209. doi: 10.1093/cercor/bhm046. [DOI] [PubMed] [Google Scholar]
  28. Ramcharan EJ, Gnadt JW, Sherman SM. The effects of saccadic eye movements on the activity of geniculate relay neurons in the monkey. Visual Neuroscience. 2001;18:253–258. doi: 10.1017/s0952523801182106. [DOI] [PubMed] [Google Scholar]
  29. Reppas JB, Usrey WM, Reid RC. Saccadic eye movements modulate visual responses in the lateral geniculate nucleus. Neuron. 2002;35:961–974. doi: 10.1016/s0896-6273(02)00823-1. [DOI] [PubMed] [Google Scholar]
  30. Ross J, Morrone MC, Goldberg ME, Burr DC. Changes in visual perception at the time of saccades. Trends in Neurosciences. 2001;24:113–121. doi: 10.1016/s0166-2236(00)01685-4. [DOI] [PubMed] [Google Scholar]
  31. Royal DW, Sary G, Schall JD, Casagrande VA. Correlates of motor planning and postsaccadic fixation in the macaque monkey lateral geniculate nucleus. Experimental Brain Research. 2006;168:62–75. doi: 10.1007/s00221-005-0093-z. [DOI] [PubMed] [Google Scholar]
  32. Sperry RW. Neural basis of the spontaneous optokinetic response produced by visual inversion. Journal of Comparative and Physiological Psychology. 1950;43:482–489. doi: 10.1037/h0055479. [DOI] [PubMed] [Google Scholar]
  33. Thilo KV, Santoro L, Walsh V, Blakemore C. The site of saccadic suppression. Nature Neuroscience. 2004;7:13–U13. doi: 10.1038/nn1171. [DOI] [PubMed] [Google Scholar]
  34. Volkmann FC, Riggs LA, White KD, Moore RK. Contrast sensitivity during saccadic eye-movements. Vision Research. 1978;18:1193–1199. doi: 10.1016/0042-6989(78)90104-9. [DOI] [PubMed] [Google Scholar]
  35. Watson TL, Krekelberg B. The relationship between saccadic suppression and perceptual stability. Current Biology. 2009;19:1040–1043. doi: 10.1016/j.cub.2009.04.052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Zhang T, Heuer HW, Britten KH. Parietal area VIP neuronal responses to heading stimuli are encoded in head-centered coordinates. Neuron. 2004;42:993–1001. doi: 10.1016/j.neuron.2004.06.008. [DOI] [PubMed] [Google Scholar]

RESOURCES