Skip to main content
NCBI home page
Journal of Neurophysiology logoLink to Journal of Neurophysiology
. 2017 Oct 18;119(1):124–133. doi: 10.1152/jn.00958.2016

Path perturbation detection tasks reduce MSTd neuronal self-movement heading responses

William K Page 1,✉, Charles J Duffy 1
PMCID: PMC5866476  PMID: 29046430

Abstract

We presented optic flow and real movement heading stimuli while recording MSTd neuronal activity. Monkeys were alternately engaged in three tasks: visual detection of optic flow heading perturbations, vestibular detection of real movement heading perturbations, and auditory detection of brief tones. Push-button RTs were fastest for tones and slower for visual and vestibular heading perturbations, suggesting that the tone detection task was easier. Neuronal heading selectivity was strongest during the tone detection task, and weaker during the visual and vestibular heading perturbation detection tasks. Heading selectivity was weaker during visual and vestibular path perturbation detection, despite our presented heading cues only in the visual and vestibular modalities. We conclude that focusing on the self-movement transients of path perturbation distracted the monkeys from their heading and reduced neuronal responsiveness to heading direction.

NEW & NOTEWORTHY Heading analysis is critical for steering and navigation. We recorded the activity of monkey cortical heading neurons during naturalistic self-movement. When the monkeys were required to respond to transient changes in their path, neuronal responses to heading direction were diminished. This suggests that the need to respond to momentary path perturbations reduces your ability to process your heading direction.

Keywords: optic flow, attention, visual motion, extrastriate, cortex

INTRODUCTION

MSTd neurons show heading-selective responses to optic flow (Bremmer et al. 2010; Duffy and Wurtz 1995) and vestibular signals about self-movement (Duffy 1998; Gu et al. 2007), making them well suited to perceptual processing for self-movement heading analysis (Heuer and Britten 2004; Kishore et al. 2012).

Task modulation of extrastriate cortical neuronal responses to visual motion have variably been found to alter response selectivity (Desimone and Duncan 1995), magnitude (Treue and Martínez Trujillo 1999), or both (Martínez Trujillo and Treue 2004). We have seen a diverse spectrum of such effects in MSTd’s optic flow responses depending on the detailed structure of the task (Dubin and Duffy 2007) and the monkey’s perceptual strategy (Page and Duffy 2008).

In MSTd, superimposed visual object and optic flow stimuli, simulating the combined cues of natural observer self-movement, evoke radically different neuronal responses depending on the behavioral task (Kishore et al. 2012). When the monkey is steering by optic flow, MSTd’s responses reflect neuronal optic flow selectivity. When the monkey is steering by object motion, the responses reflect object motion selectivity.

We have now presented circular path optic flow and real movement stimuli during push-button tasks cued by self-movement path perturbations, or by audible tones unrelated to self-movement. We find that MSTd neurons show stronger heading responses during tone detection than during path perturbation detection. We consider that path perturbation detection focuses the monkeys on self-movement transients, distracting them from their heading direction on the circular path.

METHODS

Animal preparation.

Single neurons were recorded through chronically implanted recording chambers accessing the cerebral cortex of two rhesus monkeys. Implantation surgery was preceded by sedation using ketamine (15 mg/kg im) and Robinul (0.011 mg/kg im), which was followed by venous catheterization, endotracheal intubation, oro-gastric intubation, and isoflurane general anesthesia. Heart rate, temperature, expired CO2, and the depth of anesthesia were monitored throughout the procedures described below.

Monkeys were placed in a cranial stereotaxic apparatus where the shaved scalp was washed with a betadine and draped before excising a ~3 cm × ~6 cm section of calvarial scalp. Bilateral recording cylinders were placed over 2-cm trephine holes above MSTd (AP –2 mm, ML ±15 mm, angle 0) along with a head holder and scleral search coils (Judge et al. 1980). Dental pins were implanted as anchors encased in several layers of dental acrylic to form an implant cap.

After postoperative recovery, the monkeys were returned to their home cage where monitoring guided the administration of Banamine (1 mg/kg im) analgesia under veterinary supervision. The excised scalp edge and recording chambers were gently debrided and wiped with betadine daily. All protocols were approved by the University of Rochester Committee on Animal Research and complied with U.S. Public Health Service and Society for Neuroscience policy on the care of laboratory animals.

When fully recovered from surgery, the animals were trained to maintain visual fixation that was monitored using the magnetic search coils. The monkeys were trained to push a button when hearing a tone, or seeing a directional deviation from a circular path simulated by optic flow, or feeling a directional deviation from a circular path of translational movement. The monkeys maintained centered visual fixation (±3°) in all stimulus presentation epochs. Fixation duration was increased over months of training so that fixation would be sustained throughout the circular path stimuli. The magnitude of the optic flow, sled movement, and audible tone stimuli were adjusted during training to evoke similar push-button response times for all three modalities. Correct push-button responses within 1,500 ms of behavioral stimulus onset led to a distinctly different audible tone and liquid reward and the removal of visual stimuli (i.e., optic flow and fixation target) for 4 s. Fixation breaks led to a pause in recording and a 1-s removal of visual stimuli, although any ongoing sled movement continued on the circular path. When the fixation target reappeared, the monkey had 500 ms to refixate before recording resumed.

Experimental design.

Each block of trials consisted of 10 complete circular path trials in which the selected cue was presented at one of eight nonconsecutive, randomly selected locations equally spaced on the circular path of self-movement (Fig. 1). The monkeys maintained screen-centered visual fixation throughout all recorded periods with the shape of the fixation target, indicating the task relevant cue in that block of trials: an open circle indicated that path perturbation would randomly occur in the optic flow. An open triangle indicated that path perturbation would randomly occur in the real sled movement. An open square indicated that an audible tone would randomly be sounded.

Fig. 1.

Fig. 1.

Open in a new tab

Vestibular and visual self-movement stimuli were presented during single neuron recordings. A: monkeys underwent circular path translational self-movement (red circle) either viewing an optic flow simulation of circular path self-movement on a room-sized sled, or experiencing sled movement on a circular path, or congruently combined optic flow and sled movement. The monkey was engaged in three push-button task conditions: brief path perturbations (green wavelet) from the optic flow-simulated circular path, the sled movement circular path, or a brief audible tone presented along the circular path. In all three cue modalities, the push-button cue was presented at one of eight locations around the circle. B: a table representing the seven stimulus × cue conditions studied. Three heading stimulus modalities conditions were optic flow alone, sled movement alone, and combined stimuli (abscissa). Push-button task cue modalities combined path perturbation in the presented heading stimulus plus an auditory tone detection task.

We presented seven circular path self-movement stimulus-task conditions as a series of blocked trials (Fig. 1A). Three heading stimulus conditions were used: 1) optic flow alone, 2) sled movement alone, and 3) optic flow and sled movement with the two modalities combined congruently, that is, simultaneously presenting the same heading direction (Fig. 1B, abscissa). These stimulus conditions were paired with tasks in which the monkey pressed a button to respond to a cue and earn a liquid reward (Fig. 1B, ordinate). The three stimulus conditions were presented in the order: optic flow alone, combined stimuli, and sled movement alone. This provided a smoother transition between the complex task conditions and resulted in more completed studies of optic flow alone than combined stimuli and sled movement alone. The two seemingly missing conditions are not possible.

During heading stimulus presentation by optic flow alone, the monkey completed blocks of trials in which its push-button response cue was either 1) a transient perturbation of the optic flow’s simulated circular path, or 2) an audible tone. During stimulation by sled movement alone, the monkey completed blocks of trials in which its push-button response cue was either: 1) a transient perturbation of sled movement’s circular path, or 2) an audible tone. During stimulation by congruently combined optic flow and sled movement, the monkey completed blocks of trials in which its push-button response cue was either: 1) a transient perturbation of the optic flow’s simulated circular path, 2) a transient perturbation of the sled movement’s circular path, or 3) an audible tone. The perturbation octant selected for each circular excursion is excised from the continuous data stream. Averaging across all circular excursions in that condition provides the basis for a seamless rendering of firing rate across all octants.

Optic flow stimuli.

Computer-generated optic flow stimuli were presented on a 47′′ LCD flat screen TV at 60 Hz across the central 45° × 45° (23′′ × 23′′) of the monkeys’ visual field during centered fixation. Optic flow stimuli consisted of ~2,500 single-pixel white dots (0.19° at 2.61 cd/m2) on a dark background (0.18 cd/m2). The optic flow stimuli simulated the scene during observer movement through a three-dimensional (3D) cloud of illuminated points filling a space having a depth of field extending from 1.5 to 9 m in front of the monkey. In the first frame of each optic flow display, the dots were distributed in a random pattern filtered to establish uniform dot density across the screen. Dots were removed either by passing out of the field of view or by the expiration of a randomly distributed dot duration ranging from 1 to 3 s. Dot speed was a sin2 function of dot eccentricity relative to the angle between the monkeys’ centered fixation line of sight, and each dot’s screen location relative to the current translational direction focus-of-expansion or contraction in the simulated 3D space. Each dot’s movement was adjusted to correspond to its randomly assigned depth location in the simulated 3D space.

Sled movement stimuli.

We used a two-axis, dual-rail, computer-controlled, motorized sled to present translational movement stimuli. The monkey chair, eye coil, and video display systems were mounted on a 1 × 1 m sled platform. Sled movements were controlled with position, speed, and acceleration feedback from the drive motors. That signal was sampled at 125 Hz and stored as an analog record. Stimulus presentation and experimental conditions were controlled by the PC-based real-time experimental system (Hays et al. 1982).

Each series of movement trials began with the sled at the front-center location with movement toward the monkeys’ right side assigned labeled the 0° heading direction. The platform first accelerated at 45 cm/s2 for 1.33 s and then moved on a clockwise circular path for 10 complete repetitions around the room; finally, the platform decelerated at 45 cm/s2 for 1.33 s. The sled carried the monkey on a circular (t = 10.5 s, radius = 1 m, v = 60 cm/s) path that was perturbed by the presentation of 1.5-s path perturbations that cued the monkeys’ button-press responses. The perturbation direction was always toward the center of the circular path, resulting in different perturbation directions at each octant interval.

The sled motion was controlled with a Parker ACR9000 and two Aries servo drives with analog interface to the ACR9000. The circular path was created by executing a “SINE X(i,j,k,l) SINE Y(i,j-90,k,l)” function where i = center position, j = 90 (phase), k = 3,600 degrees of movement (10 excursions around the circle), l = 1000 (1-m radius in mm). When a path perturbation was triggered, a “JOG X z1, JOG Y z2” command was executed, where sqrt[(z1^2) + (z2^2)] = 40 mm. This added a 40-mm jog to the circular path movement. After 1.0 s, the jog movement was reversed to return the sled to the circular path. The values for z1 and z2 varied depending on which of the eight octants the path deviation occurred. At all octants, the initial jog direction was toward the center of the circular path (Fig. 1A, green line).

In training, the size of the sled movement perturbations was adjusted to obtain 80% successful trials. Optic flow perturbations were adjusted to be of a corresponding direction, duration, and magnitude, simulating the sled movement perturbations. Tone cues were presented for 35 ms at 78 dB, 18 dB above the continuous ambient noise due to air flow in the room HVAC system. Both the optic flow perturbations and tones yielded nearly 100% correct trials in both monkeys. In the sled movement-alone condition, sled movement occurred with only the visual fixation point on the screen. In the combined stimulus condition, sled movement was accompanied by optic flow stimuli, simulating their naturalistically congruent combination.

Single neuron recording.

Epoxy-coated tungsten microelectrodes were passed through transdural guide tubes mounted in a cylindrical recording chamber grid (Crist et al. 1988). Neuronal activity was monitored as the electrode was advanced to identify gray and white matter physiological landmarks. On isolating the activity of a single neuron, its receptive field was mapped using computer-generated local motion patches, while the monkey fixated a screen-centered target. Gray matter layers at stereotaxic locations consistent with MSTd, that include neurons with large (>20°), direction-selective, pattern motion-preferring receptive fields, are considered to be part of MSTd (Duffy and Wurtz 1991, 1995; Komatsu and Wurtz 1988a). Single neuron discharges were isolated using a commercial algorithm (Alpha Omega, Atlanta, GA). These events were digitized into a continuous record of neuronal discharge times that were stored with stimulus and behavioral event markers for off-line analysis in MatLab (Mathworks, Natick, MA).

Recording sites.

When recordings were completed, we placed electrolytic marks (25 μA × 25 s) at selected grid positions and depths. Prior to euthanasia, one of these monkeys died unexpectedly. The other monkey underwent pentobarbital euthanasia with perfusion by heparinized saline followed by formalin. The brain was removed for formalin fixation, with posterior whole brain sectioning in 50-μm sections with every fourth section Nissl stained for cell body layers and every fifth section Luxol Fast Blue stained for myelinated fibers. Examination of stained sections confirmed neuronal recording in the anterior bank of the superior temporal sulcus within area MSTd (Komatsu and Wurtz 1988b).

Data analysis.

Neuronal spike times were convolved with a Gaussian function having a sigma of 25 ms to produce spike density response functions averaging the 10 stimulus trials in each stimulus-task paradigm. Task-related cue presentation intervals of 1.5 s were excised from all trials and then averaged to create spike densities plotted and sampled with reference to circular path-heading direction. We derived sample responses across neurons and stimulus paradigms. The responses of each neuron were normalized to the peak average firing rate during the tone detection task in the optic flow-alone paradigm. Normalized responses were averaged across neurons by aligning on the peak response in each condition of each stimulus paradigm. Data were statistically analyzed using the measures described in the text (SPSS 2013), but for the circular statistics that were implemented in MatLab to support heading selectivity and response profile analyses.

RESULTS

Push-button response times.

Circular paths of self-movement present a complete series of heading directions in the ground plane (Fig. 1A). We presented circular path stimuli using optic flow, sled movement, or both. These stimuli were combined with a series of seven push-button stimulus-task conditions (see methods and Fig. 1B). The magnitude of the path perturbations and tones was adjusted so that both monkeys responded within the 1.5-s response window applied to all three modalities: optic flow or sled movement path perturbations or the audible tone task cues (Fig. 2).

Fig. 2.

Fig. 2.

Open in a new tab

Comparison of the response times for the various stimulus and task conditions in the two monkeys. The response times (ordinate) for the two monkeys (A: M606; B: M125; C: both) with the optic flow cue (red), translational movement cue (green), and tone cue (blue) stimuli are shown for each of three stimulus conditions (abscissa): optic flow alone (left), real movement alone (middle), and both optic flow and real movement (right). The reaction times were shortest for the tone-cued trials and longest for the sled movement-cued trials.

Push-button response times (RTs) were compared in a three-way ANOVA for stimulus condition (seven), path heading (eight directions), and cue interval (eight intervals, 2nd–9th heading segment since the previous cue). All main and interaction effects were significant: condition (F6,5364 = 108.06, P < 0.001; Tukey honestly significant differences (THSDs): sled > flow > tone), heading (F7,5364 = 5.13, P < 0.001; THSDs: forward and backward > right and left), interval (F8,5364 = 3.61, P < 0.001; THSDs: first interval since previous perturbation cue > last), condition × location (F42,5364 = 2.62, P < 0.001), condition × interval (F47,5364 = 1.67, P = 0.003), location × interval (F56,5364 = 1.50, P = 0.01), condition × location × interval (F293,5364 = 1.51, P < 0.001). The push-button hits, misses, and false positives mirror the RT effects (Table 1). These effects suggest that significant stimulus and task effects interact to affect RTs across heading conditions and self-movement locations.

Table 1.

Perturbation detection push-button responses

Condition and Task Monkey Hits Misses False Positives
Flow vs. tone
    Flow M125 93.5% 2.6% 3.9%
M606 90.3% 3.1% 6.5%
Both 92.3% 2.8% 5.0%
    Tone M125 98.8% 1.0% 0.2%
M606 99.6% 0.2% 0.2%
Both 99.2% 0.7% 0.2%
Movement vs. tone
    Movement M125 75.6% 1.9% 22.6%
M606 49.4% 12.0% 38.7%
Both 65.0% 5.9% 29.0%
    Tone M125 96.1% 0.0% 3.9%
M606 94.3% 1.1% 4.6%
Both 95.4% 0.5% 4.2%
Flow vs. movement vs. tone
    Flow M125 89.8% 0.8% 9.4%
M606 91.8% 2.9% 5.3%
Both 90.6% 1.6% 7.7%
    Movement M125 77.4% 2.1% 20.5%
M606 50.6% 18.8% 30.7%
Both 66.8% 8.7% 24.5%
    Tone M125 97.0% 1.4% 1.6%
M606 97.0% 0.6% 2.5%
Both 97.0% 1.1% 1.9%
Open in a new tab

The push-button response performance of the monkeys reflects the RT data in Fig. 2. Tone responses are fastest and most accurate. Movement responses are slowest and least accurate. Optic flow responses are intermediate.

Task effects on single-neuron responses.

We recorded the activity of 128 neurons from three hemispheres of two adult Rhesus monkeys. All neurons yielding significant effects of circular path heading direction, task condition, or their interactions (two-way ANOVA P < 0.05), were included in further analyses. Responses in the optic flow alone condition yielded significant effects in 85% (109/128) of the neurons. Responses in the sled movement-alone condition yielded significant effects in 57% (28/49) of the neurons. Responses to combined optic flow and sled movement condition yielded significant effects in 80% (45/56) of the neurons.

MSTd neurons showed task-related modulation of their heading responses that reflect the effects of optic flow and sled movement circular path stimuli, as the tone does not impart heading information. Response modality selectivity can be seen as responsiveness to either optic flow or sled movement stimuli (Fig. 3, A and B). However, the lack of activation by a heading stimulus (e.g., sled movement in Fig. 3B) does not necessarily indicate the absence of effects on neuronal responsiveness, and those effects can be task dependent. For example, there is no net activation when activating optic flow is combined with nonactivating sled movement during the sled movement path perturbation task (Fig. 3C). In that case, the sled movement path perturbation task appears to suppress the simultaneously presented optic flow responses, yielding the apparent nonresponsiveness seen when sled movement was presented alone.

Fig. 3.

Fig. 3.

Open in a new tab

Activity of an MSTd neuron (firing rate, ordinate) recorded during circular path stimulus presentation (10.5 s, abscissa). The ±750-ms interval of cueing by path perturbations or audible tones was excluded from the records of each trial. A: activity evoked by optic flow circular path simulation is larger during the audible tone detection task (blue) than during the optic flow path perturbation detection task (red). B: activity evoked by translational movement circular path stimulation is similar during the audible tone detection task (blue) and the translational movement path perturbation detection task (green). C: activity evoked by congruently combined optic flow and translational movement circular path stimuli is largest during the audible tone detection task (blue), smaller during the optic flow path perturbation detection task (red), and greatly diminished during translational movement path perturbation detection task (green).

Across neurons, there is great diversity in self-movement and task effects. Responses to optic flow presented alone showed changes concentrated at their preferred heading directions with relative enhancement (Fig. 4A, left) or suppression (Fig. 4A, right) during the tone detection task. Circular path sled movement presented alone showed a similar range of task effects, typically with broader, lower-amplitude, heading selectivities (Fig. 4B). Responses to combined optic flow and sled movement showed response diversity within single responses, with heading-selective (Fig. 4C, left) or nonselective (Fig. 4C, right) effects of task on segments of the heading response profiles.

Fig. 4.

Fig. 4.

Open in a new tab

Task effects on heading-selective responses to circular path stimuli. A: neuronal activity (ordinate) throughout the optic flow circular path simulation (10.5 s, abscissa) during push-button cueing by optic flow path perturbation (red) or by an audible tone (blue). Left: a neuron that shows larger optic flow heading responses during the tone detection task. Right: a neuron that shows larger optic flow heading responses during the optic flow path perturbation detection task. B: task effects on heading-selective responses to sled movement circular path stimuli (format as in A). Left: a neuron that shows larger sled movement heading responses during the tone detection task. Right: a neuron that shows larger sled movement heading responses during the sled movement path perturbation detection task. C: task effects on heading-selective responses to combined optic flow and translational movement circular path stimuli (format as in A). Left: a neuron that shows a shift of heading selectivity with a larger response during the tone detection task. Right: a neuron that shows larger heading responses during the optic flow path perturbation detection task.

The effects seen in the Tone vs. Flow vs. Movement condition were further analyzed using two-way ANOVAs with main effects of heading (eight directions), task condition (three tasks), and their interactions. Across both monkeys, all three factors are significant: direction (F7,3280 = 69.316, P < 0.001), condition (F2,3280 = 6.743, P = 0.001), and interaction (F14,3280 = 1.762, P = 0.038). Individually, both monkeys showed significant main effects: Monkey 606 (condition F2,1727 = 4.164, P = 0.016; direction F7,1727 = 33.964, P < 0.001; interaction F14,1727 = 1.121, P = 0.338). Monkey 125 (direction (F7,1551 = 36.816, P < 0.001; condition F2,1551 = 4.198, P = 0.015, interaction F14,1551 = 0.793, P = 0.677).

We examined single neuron responses using two-way ANOVAs with main effects of heading (eight directions) and task condition (two or three tasks) to identify neurons with significant task effects or direction by task interactions. This approach was applied to all three stimulus conditions: optic flow alone with the optic flow or tone tasks, sled movement alone with the sled movement or tone tasks, and combined stimuli with the optic flow, sled movement, or tone tasks.

In the optic flow-alone paradigm, 84% (108/128) of the neurons yield significant effects: main effects of heading direction (42%, 54/128), main effects of task condition (4%, 5/128), both main effects (25%, 32/128), or their interactions (13%, 17/108). Equal numbers of neurons show larger responses in the tone detection task (50%, 54/108, Fig. 4A, left) and the optic flow path perturbation task (50%, 54/108, Fig. 4A, right).

In the sled movement-alone paradigm, 57% (28/49) of the neurons yield significant main effects of heading direction (6%, 3/49), or task condition (31%, 15/49), both (10%, 5/49), or their interactions (10%, 5/49). More of these neurons showed larger responses during tone detection (61%, 17/28, Fig. 4B, left) than during sled movement path perturbation (39%, 11/28, Fig. 4B, right).

In the combined optic flow and sled movement paradigm, 83% (45/54) of the neurons yield significant main effects of heading direction (15%, 8/54), or task condition (17%, 9/54), both (30%, 16/54), or their interactions (22%, 12/54). The tone detection task yielded larger responses (42%, 19/45, Fig. 4C, left), in more neurons than either the optic flow path perturbation (29%, 13/45, Fig. 4C, right), or the sled movement path perturbation tasks (29%, 13/45, Fig. 4C, left).

These effects illustrate the complexity of MSTd neuronal heading selectivity’s interactions with the monkey’s ongoing behavioral tasks.

Neuronal response profiles.

This classification of task effects is complicated by many neuron’s showing nonunimodal heading direction response profiles, most commonly bimodal (Fig. 4C, left), as previously seen in related studies of MSTd neurons (Sato et al. 2013). We classify response profiles by the heading difference between their largest and second largest responses. Neurons with the largest and second largest responses within 90° are considered unimodal, those beyond 90° are considered nonunimodal. The comparison of averaged unimodal and nonunimodal responses for each stimulus-task condition shows the tendency to bimodality in the nonunimodal responses (Fig. 5, left).

Fig. 5.

Fig. 5.

Open in a new tab

The bimodality of response profiles in the sample of neurons is illustrated for the three stimulus conditions: optic flow alone (A), sled movement alone (B), and combined stimuli (C). Left: neurons were split into two groups depending on whether the second peak was close (solid lines) or further away (dashed lines) from the preferred stimulus direction. The neurons with a distant second peak have a distinct second peak for all conditions. Right: scatterplots of the degree of unimodal selectivity (abscissa) vs. bimodal selectivity (ordinate) from Rayleigh z scores for the three stimulus conditions (A–C). There is a broad continuum of the relative strengths of unimodality and bimodality in all conditions. Illustrative examples are shown in the inset responses, and their position in the scatterplots is indicated by the nearer stars.

We used circular statistical analysis to assess heading direction selectivity. The Rayleigh z measured the strength of unimodal heading selectivity, applying angle doubling to create a comparable measure of bimodality. Scatterplots of unimodal vs. bimodal directionality for each response profile show the most bimodality in the flow vs. tone conditions, and least in the movement vs. tone condition (Fig. 5, right). We used a Rayleigh P < 0.05 to classify response profiles as to their significant unimodality, bimodality, both, or neither (Table 2). This tabulation shows that mixed unimodality and bimodality is the most common profile, with unimodality alone being more common than bimodality alone, potentially suggesting both functional diversity and mechanistic complexity.

Table 2.

Response profile bimodality and unimodality

Condition and Task Unimodal Bimodal Both Neither
Flow vs. tone
    Flow 22% (24/109) 7% (8/109) 63% (69/109) 7% (8/109)
    Tone 23% (25/109) 6% (6/109) 68% (47/109) 4% (4/109)
Movement vs. tone
    Movement 18% (5/28) 21% (6/28) 43% (12/28) 18% (5/28)
    Tone 11% (3/28) 21% (6/28) 57% (16/28) 11% (3/28)
Flow vs. movement vs. tone
    Flow 20% (9/45) 2% (1/45) 64% (29/45) 13% (6/45)
    Movement 29% (13/45) 18% (8/45) 47% (21/45) 7% (3/45)
    Tone 9% (4/45) 16% (7/45) 76% (34/45) 0% (0/45)
Open in a new tab

The continuum of unimodal and bimodal response profiles is shown by the number of neurons that showed significant unimodal or bimodal Rayleigh z scores (P < 0.05). In all stimulus and task conditions (rows), the most common pattern of significant effects (columns) was in both unimodal and bimodal tests (bolded values).

The interval of path perturbation in each circular path was edited out of the continuous data stream. We examined activity during the path perturbation intervals but could not identify any substantial deviations from that neuron’s range of firing rates. As might be expected, perturbations from preferred headings tended to decrease activity, and perturbations from nonpreferred headings tended to increase activity. As path perturbation was only intended for behavioral control, there were only one or two perturbations at each path segment in each condition for each neuron. This sparse data set does not support statistical analyses.

Task effects on population responses.

Population responses in the optic flow-alone stimulus condition reveal a slightly larger peak amplitude in the tone task without significant differences (Wilk’s lambda from repeated-measures ANOVA for task effects F1,863 = 0.71, P = 0.79; Fig. 6A). Gaussian fit comparisons between these task conditions are not reported because of the nonsignificant ANOVA. In contrast, population responses in the sled movement-alone stimulus condition show significantly larger responses in the tone detection task across all heading directions (F1,223 = 15.71, P < 0.001; Fig. 6B). Comparison of Gaussian fit responses in the sled movement perturbation and tone tasks reveals higher baseline activity in the tone task (18.5%, P = 0.02) without other differences.

Fig. 6.

Fig. 6.

Open in a new tab

Effects of cue modality (line colors) on neuronal population-averaged responses (ordinates) across heading directions (abscissas). A: averaged population responses for heading directions simulated by optic flow do not show significant differences between the optic flow path perturbation detection task vs. the audible tone detection task. B: averaged population responses for heading directions in translational movement do not show significant differences between the translational movement path perturbation detection task vs. the audible tone detection task. C: averaged population responses for heading directions in combined optic flow and translational movement show significantly greater heading selectivity during the audible tone detection task than during the optic flow or translational movement path perturbation detection task. D: scatterplot of differences in averaged single neuron activity in the tone detection task minus the translational movement path perturbation detection task (ordinate), vs. in the tone detection task minus the optic flow path perturbation detection task (abscissa). The comparability of effects in the tone detection task, compared with the translational movement and optic flow path perturbation detection tasks, show similar degrees of task modulation in the visual and vestibular domains.

Sample responses to combined optic flow and sled movement yields significantly larger responses in the tone detection task (F2,357 = 5.10, P = 0.007; Fig. 6C). Pairwise comparisons between task conditions show significant differences between the optic flow and tone detection tasks (F1,358 = 9.59, P = 0.002) and between the sled movement and tone detection tasks (F1,358 = 7.72, P = 0.006), but not between the optic flow and sled movement tasks (F1,359 = 0.60, P = 0.44). Comparison of Gaussian fits to the aligned responses shows the tone task yields larger amplitude than the flow task (14.4%, P = 0.003) and a larger baseline than the movement task (11.6%, P = 0.04) without other differences.

The combined stimulus paradigm allows us to compare a single neuron’s heading responses during the tone task, to that neuron’s heading responses during the optic flow task and during the translational movement task. In individual neurons, we find similar effects in comparing tone task responses to both optic flow task responses and that neuron’s translational movement task responses (R2 = 0.49, P < 0.0001; Fig. 6D). Thus, individual neurons show the same task effects on combined optic flow and translational heading responses, whether the task was optic flow path perturbation detection or translational movement path perturbation detection.

DISCUSSION

Behavioral response effects.

We presented circular path self-movement stimuli to monkeys while they detected optic flow or sled movement path perturbations, or audible tones unrelated to the circular path (Fig. 1). The monkeys’ push-button response times varied with the task with successively faster RTs in the sled path perturbation, optic flow path perturbation, and tone detections tasks (Fig. 2). This suggests that the tone task was less demanding than the path perturbation tasks, with path perturbation placing greater attentional demands on both monkeys.

Tone or perturbation cues were equally distributed among the eight heading direction intervals on each excursion around the circular path, with every circular path containing a single cue. This resulted in the increasing probability of a cue in the upcoming heading segment if it had not occurred in the prior segments of that circular path. This is reflected in the monkeys having longer RTs to the first potential cue interval, and shorter RTs to the last, consistent with the relationship between stimulus probability and simple RT (Gordon 1967; Klemmer 1956).

There was also a significant effect of heading direction on RTs, with forward and backward headings yielding longer RTs than leftward or rightward headings. We consider this observation to be consistent with greater human perceptual (Crowell and Banks 1993) and monkey neuronal (Angelaki et al. 2011) sensitivities to forward and backward headings. That greater sensitivity may impart greater distraction from the path perturbation and tone detection tasks and result in the longer RTs during forward and backward self-movement headings.

Task effects on heading response profiles.

Task effects on neuronal responses varied across stimulus conditions. Comparing the path perturbation and tone tasks showed significant effects in 42% of the neurons in the optic flow-alone condition (Fig. 4A), 51% in the sled movement-alone condition (Fig. 4B), and 69% in the combined optic flow and sled movement condition (Fig. 4C). Thus, task effects on heading responses were greater with combined stimuli. We consider that the monkeys’ heading processing may be less distracted by the easier tone task than by the more difficult path perturbation tasks. This effect may be enhanced by the monkeys’ greater effort to ignore the more naturalistically compelling combined stimuli to maintain performance on perturbation detection.

These stimulus-task conditions elicited diverse heading direction responses with neurons varying from unimodal, to mixed, to bimodal response profiles (Figs. 3–5). There are prominent bimodal response profiles in all stimulus-task conditions (Fig. 5 and Table 2). Double-peaked bimodal MSTd responses have been seen across stimulus compositions in optic flow heading perception studies (Chen et al. 2011), across task conditions in working memory (Sato et al. 2013), and in naturalistic steering studies (Jacob and Duffy 2015).

We consider that MSTd neurons’ response bimodality may derive from two sources: Bimodal responses to optic flow may reflect optic flow’s radial pattern of locally planar motion directions, which interact with the pattern of planar motion directional subfields of MSTd neuron’s large receptive fields (Yu et al. 2010).

These effects might also be viewed as component responses to full-field stimulation (Khawaja et al. 2009; Movshon et al. 1986; Richert et al. 2013), possibly promoted by task effects. Bimodal responses to sled movement may reflect the inertial equivalence of acceleration in one direction and deceleration in the opposite direction, which may be highlighted by the large-scale circular motion paths of our studies (Duffy 1998; Gu et al. 2007; Schlack et al. 2002). Combined vestibular and visual stimuli might generate the observed spectrum of composite response profiles.

The circular paths may also account for the relatively smaller sled movement responses seen in these studies compared with earlier work (Duffy 1996; Gu et al. 2006), either because of the continually changing heading, cortical habituation to the sustained presentation of the same circular path, or possibly because task effects disrupted vestibular responsiveness. Distinguishing among such factors is beyond the scope of this study.

Naturalistically combined stimuli.

Averaged responses to optic flow or sled movement alone show only small net differences between the tone detection task and the path perturbation task (Fig. 6, A and B). In contrast, averaged responses during the combined presentation of heading congruent optic flow and sled movement show more substantial differences across task conditions. Peak responses in the tone task are larger than peak responses in either the optic flow or sled movement path perturbation tasks. However, peak responses to optic flow and sled movement path perturbation yield very similar responses (Fig. 6C).

With combined stimuli, similar effects of the optic flow and sled movement path perturbation tasks are also seen in single neurons: Neurons that show large differences between responses during tone detection and optic flow path perturbation, also show large differences between responses during tone detection and sled movement path perturbation (Fig. 6D). This suggests that these effects reflect single-neuron response profile sensitivity to differences between the tone task and the path perturbation task, rather than specific sensitivity to either the optic flow or sled movement path perturbation tasks.

Previously, we have seen MSTd neuronal optic flow responses increased in optic flow-guided tasks, with smaller optic flow responses in object-guided tasks (Dubin and Duffy 2007). In that study, the optic flow heading was the task-related cue, and stronger heading responses were obtained. In the current study, when path perturbation unrelated to heading guided the tasks, weaker heading responses were obtained. Thus, when heading in optic flow guides the task, heading responses are stronger. Whereas, when a nonheading cue in optic flow guides the task, heading responses are weaker.

These results may be linked to our previous finding that steering by the global heading in optic flow increases MSTd neuronal responses, whereas steering by the local motion cues in optic flow decreases optic flow responses (Page and Duffy 2008). Similarly, we consider that our path perturbation detection tasks focused the monkeys on local motion transients in the optic flow, distracting them from the optic flow’s heading direction. The easier tone detection task may have been less distracting and did not redirect self-movement analysis from the heading direction, which led to the tone task’s stronger heading responses.

Implications for distracted driving.

Both the tone and path perturbation tasks were unrelated to heading direction, but only the path perturbation task placed demands on visual and vestibular self-movement processing. Those demands focused the monkeys on the detection of brief transients in the rate of heading change, potentially distracting the monkeys from their current heading on the circular path. Thus, path perturbation detection may have distracted the monkeys from heading processing, and, hence, may have reduced neuronal heading responses.

We recognize a potential functional dichotomy between self-movement analysis for detecting path perturbations vs. self-movement analysis for heading estimation. In an optic flow-delayed steer to sample task, we presented a naturalistic sequence of heading estimation, deviation detection, and corrective steering back to the initial heading. MSTd neuronal activity varied across the phase of this task from global optic flow analysis, to local motion analysis, to spatial location analysis (Jacob and Duffy 2015). This suggests that immediate behavioral task demands can alter MSTd neuronal activity across a continuum from greater emphasis on the heading direction in optic flow, to its pattern of local motion, to the spatial location of the center of radial motion.

Alternatively, we might infer some degree of functional antagonism between steering’s focus on instantaneous self-movement and navigation’s focus on sustained paths. This view recognizes that path perturbation detection and heading direction estimation may induce dual-task interference, reflecting a capacity-sharing conflict (Fagot and Pashler 1992; Pashler and OʼBrien 1993) in MSTd. Such dual task interference may occur when two tasks rely on overlapping cortical resources (Khoe et al. 2006; Nash and Fernandez 1996), or engage competitive interactions between cortical areas (Marcantoni et al. 2003; Matthews et al. 2006), potentially under top-down attentional control (Hiraga et al. 2009; Watanabe and Funahashi 2014). The impact of competing mechanisms, and the psychological refractory period induced by task switching, is evident in attentional influences on driving performance (Hibberd et al. 2013) and may underlie the common partitioning of steering and navigating in demanding circumstances (e.g., pilots and navigators).

GRANTS

This work was supported by National Eye Institute Grants R01-EY022062 and Office of Naval Research Grant N000141110525.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

W.K.P. and C.J.D. conceived and designed research; W.K.P. performed experiments; W.K.P. and C.J.D. analyzed data; W.K.P. and C.J.D. interpreted results of experiments; W.K.P. and C.J.D. prepared figures; W.K.P. and C.J.D. drafted manuscript; W.K.P. and C.J.D. edited and revised manuscript; W.K.P. and C.J.D. approved final version of manuscript.

ACKNOWLEDGMENTS

We gratefully acknowledge the software and computer systems assistance of William Vaughn and the animal care and training assistance of Sherry Estes.

REFERENCES

  1. Angelaki DE, Gu Y, Deangelis GC. Visual and vestibular cue integration for heading perception in extrastriate visual cortex. J Physiol 589: 825–833, 2011. doi: 10.1113/jphysiol.2010.194720. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bremmer F, Kubischik M, Pekel M, Hoffmann KP, Lappe M. Visual selectivity for heading in monkey area MST. Exp Brain Res 200: 51–60, 2010. doi: 10.1007/s00221-009-1990-3. [DOI] [PubMed] [Google Scholar]
  3. Chen A, DeAngelis GC, Angelaki DE. A comparison of vestibular spatiotemporal tuning in macaque parietoinsular vestibular cortex, ventral intraparietal area, and medial superior temporal area. J Neurosci 31: 3082–3094, 2011. doi: 10.1523/JNEUROSCI.4476-10.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Crist CF, Yamasaki DS, Komatsu H, Wurtz RH. A grid system and a microsyringe for single cell recording. J Neurosci Methods 26: 117–122, 1988. doi: 10.1016/0165-0270(88)90160-4. [DOI] [PubMed] [Google Scholar]
  5. Crowell JA, Banks MS. Perceiving heading with different retinal regions and types of optic flow. Percept Psychophys 53: 325–337, 1993. doi: 10.3758/BF03205187. [DOI] [PubMed] [Google Scholar]
  6. Desimone R, Duncan J. Neural mechanisms of selective visual attention. Annu Rev Neurosci 18: 193–222, 1995. doi: 10.1146/annurev.ne.18.030195.001205. [DOI] [PubMed] [Google Scholar]
  7. Dubin MJ, Duffy CJ. Behavioral influences on cortical neuronal responses to optic flow. Cereb Cortex 17: 1722–1732, 2007. doi: 10.1093/cercor/bhl083. [DOI] [PubMed] [Google Scholar]
  8. Duffy CJ. Real movement responses of optic flow neurons in MST. Soc Neurosci Abstr 22: 1692, 1996. [Google Scholar]
  9. Duffy CJ. MST neurons respond to optic flow and translational movement. J Neurophysiol 80: 1816–1827, 1998. [DOI] [PubMed] [Google Scholar]
  10. Duffy CJ, Wurtz RH. Sensitivity of MST neurons to optic flow stimuli. I. A continuum of response selectivity to large-field stimuli. J Neurophysiol 65: 1329–1345, 1991. [DOI] [PubMed] [Google Scholar]
  11. Duffy CJ, Wurtz RH. Response of monkey MST neurons to optic flow stimuli with shifted centers of motion. J Neurosci 15: 5192–5208, 1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fagot C, Pashler H. Making two responses to a single object: implications for the central attentional bottleneck. J Exp Psychol Hum Percept Perform 18: 1058–1079, 1992. doi: 10.1037/0096-1523.18.4.1058. [DOI] [PubMed] [Google Scholar]
  13. Gordon IE. Stimulus probability and simple reaction time. Nature 215: 895–896, 1967. doi: 10.1038/215895a0. [DOI] [PubMed] [Google Scholar]
  14. Gu Y, DeAngelis GC, Angelaki DE. A functional link between area MSTd and heading perception based on vestibular signals. Nat Neurosci 10: 1038–1047, 2007. doi: 10.1038/nn1935. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gu Y, Watkins PV, Angelaki DE, DeAngelis GC. Visual and nonvisual contributions to three-dimensional heading selectivity in the medial superior temporal area. J Neurosci 26: 73–85, 2006. doi: 10.1523/JNEUROSCI.2356-05.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hays AV. A UNIX-based multiple process system for real-time data acquisition and control. WESCON Conference Proceedings 2: 1–10, 1982. [Google Scholar]
  17. Heuer HW, Britten KH. Optic flow signals in extrastriate area MST: comparison of perceptual and neuronal sensitivity. J Neurophysiol 91: 1314–1326, 2004. doi: 10.1152/jn.00637.2003. [DOI] [PubMed] [Google Scholar]
  18. Hibberd DL, Jamson SL, Carsten OM. Mitigating the effects of in-vehicle distractions through use of the Psychological Refractory Period paradigm. Accid Anal Prev 50: 1096–1103, 2013. doi: 10.1016/j.aap.2012.08.016. [DOI] [PubMed] [Google Scholar]
  19. Hiraga CY, Garry MI, Carson RG, Summers JJ. Dual-task interference: attentional and neurophysiological influences. Behav Brain Res 205: 10–18, 2009. doi: 10.1016/j.bbr.2009.07.019. [DOI] [PubMed] [Google Scholar]
  20. Jacob MS, Duffy CJ. Steering transforms the cortical representation of self-movement from direction to destination. J Neurosci 35: 16055–16063, 2015. doi: 10.1523/JNEUROSCI.2368-15.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Judge SJ, Richmond BJ, Chu FC. Implantation of magnetic search coils for measurement of eye position: an improved method. Vision Res 20: 535–538, 1980. doi: 10.1016/0042-6989(80)90128-5. [DOI] [PubMed] [Google Scholar]
  22. Khawaja FA, Tsui JM, Pack CC. Pattern motion selectivity of spiking outputs and local field potentials in macaque visual cortex. J Neurosci 29: 13702–13709, 2009. doi: 10.1523/JNEUROSCI.2844-09.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Khoe W, Freeman E, Woldorff MG, Mangun GR. Interactions between attention and perceptual grouping in human visual cortex. Brain Res 1078: 101–111, 2006. doi: 10.1016/j.brainres.2005.12.083. [DOI] [PubMed] [Google Scholar]
  24. Kishore S, Hornick N, Sato N, Page WK, Duffy CJ. Driving strategy alters neuronal responses to self-movement: cortical mechanisms of distracted driving. Cereb Cortex 22: 201–208, 2012. doi: 10.1093/cercor/bhr115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Klemmer ET. Time uncertainty in simple reaction time. J Exp Psychol 51: 179–184, 1956. doi: 10.1037/h0042317. [DOI] [PubMed] [Google Scholar]
  26. Komatsu H, Wurtz RH. Relation of cortical areas MT and MST to pursuit eye movements. I. Localization and visual properties of neurons. J Neurophysiol 60: 580–603, 1988a. [DOI] [PubMed] [Google Scholar]
  27. Komatsu H, Wurtz RH. Relation of cortical areas MT and MST to pursuit eye movements. III. Interaction with full-field visual stimulation. J Neurophysiol 60: 621–644, 1988b. [DOI] [PubMed] [Google Scholar]
  28. Marcantoni WS, Lepage M, Beaudoin G, Bourgouin P, Richer F. Neural correlates of dual task interference in rapid visual streams: an fMRI study. Brain Cogn 53: 318–321, 2003. doi: 10.1016/S0278-2626(03)00134-9. [DOI] [PubMed] [Google Scholar]
  29. Martínez-Trujillo JC, Treue S. Feature-based attention increases the selectivity of population responses in primate visual cortex. Curr Biol 14: 744–751, 2004. doi: 10.1016/j.cub.2004.04.028. [DOI] [PubMed] [Google Scholar]
  30. Matthews A, Garry MI, Martin F, Summers J. Neural correlates of performance trade-offs and dual-task interference in bimanual coordination: an ERP investigation. Neurosci Lett 400: 172–176, 2006. doi: 10.1016/j.neulet.2006.02.043. [DOI] [PubMed] [Google Scholar]
  31. Movshon JA, Adelson EH, Gizzi MS, Newsome WT. The analysis of moving visual patterns. Exp Brain Res 11: 117–151, 1986. doi: 10.1007/978-3-662-09224-8_7. [DOI] [Google Scholar]
  32. Nash AJ, Fernandez M. P300 and allocation of attention in dual-tasks. Int J Psychophysiol 23: 171–180, 1996. doi: 10.1016/S0167-8760(96)00049-9. [DOI] [PubMed] [Google Scholar]
  33. Page WK, Duffy CJ. Cortical neuronal responses to optic flow are shaped by visual strategies for steering. Cereb Cortex 18: 727–739, 2008. doi: 10.1093/cercor/bhm109. [DOI] [PubMed] [Google Scholar]
  34. Pashler H, O’Brien S. Dual-task interference and the cerebral hemispheres. J Exp Psychol Hum Percept Perform 19: 315–330, 1993. doi: 10.1037/0096-1523.19.2.315. [DOI] [PubMed] [Google Scholar]
  35. Richert M, Albright TD, Krekelberg B. The complex structure of receptive fields in the middle temporal area. Front Syst Neurosci 7: 2, 2013. doi: 10.3389/fnsys.2013.00002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sato N, Page WK, Duffy CJ. Task contingencies and perceptual strategies shape behavioral effects on neuronal response profiles. J Neurophysiol 109: 546–556, 2013. doi: 10.1152/jn.00377.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Schlack A, Hoffmann KP, Bremmer F. Interaction of linear vestibular and visual stimulation in the macaque ventral intraparietal area (VIP). Eur J Neurosci 16: 1877–1886, 2002. doi: 10.1046/j.1460-9568.2002.02251.x. [DOI] [PubMed] [Google Scholar]
  38. SPSS IBM SPSS Statistics for Windows, version 22.0 Armonk, NY: IBM, 2013. [Google Scholar]
  39. Treue S, Martínez Trujillo JC. Feature-based attention influences motion processing gain in macaque visual cortex. Nature 399: 575–579, 1999. doi: 10.1038/21176. [DOI] [PubMed] [Google Scholar]
  40. Watanabe K, Funahashi S. Neural mechanisms of dual-task interference and cognitive capacity limitation in the prefrontal cortex. Nat Neurosci 17: 601–611, 2014. doi: 10.1038/nn.3667. [DOI] [PubMed] [Google Scholar]
  41. Yu CP, Page WK, Gaborski R, Duffy CJ. Receptive field dynamics underlying MST neuronal optic flow selectivity. J Neurophysiol 103: 2794–2807, 2010. doi: 10.1152/jn.01085.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Journal of Neurophysiology are provided here courtesy of American Physiological Society

RESOURCES