SEARCH

SEARCH BY CITATION

Keywords:

  • face;
  • face-like patterns;
  • monkey;
  • pulvinar;
  • subcortical pathway

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

The pulvinar nuclei appear to function as the subcortical visual pathway that bypasses the striate cortex, rapidly processing coarse facial information. We investigated responses from monkey pulvinar neurons during a delayed non-matching-to-sample task, in which monkeys were required to discriminate five categories of visual stimuli [photos of faces with different gaze directions, line drawings of faces, face-like patterns (three dark blobs on a bright oval), eye-like patterns and simple geometric patterns]. Of 401 neurons recorded, 165 neurons responded differentially to the visual stimuli. These visual responses were suppressed by scrambling the images. Although these neurons exhibited a broad response latency distribution, face-like patterns elicited responses with the shortest latencies (approximately 50 ms). Multidimensional scaling analysis indicated that the pulvinar neurons could specifically encode face-like patterns during the first 50-ms period after stimulus onset and classify the stimuli into one of the five different categories during the next 50-ms period. The amount of stimulus information conveyed by the pulvinar neurons and the number of stimulus-differentiating neurons were consistently higher during the second 50-ms period than during the first 50-ms period. These results suggest that responsiveness to face-like patterns during the first 50-ms period might be attributed to ascending inputs from the superior colliculus or the retina, while responsiveness to the five different stimulus categories during the second 50-ms period might be mediated by descending inputs from cortical regions. These findings provide neurophysiological evidence for pulvinar involvement in social cognition and, specifically, rapid coarse facial information processing.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

The pulvinar nuclei are located in the posterior region of the thalamus and are proportionally larger in higher mammals, such as primates, having the largest dimensions in the human brain (Browne & Simmons, 1984). The pulvinar receives visual inputs from subcortical structures, including the superficial and deep layers of the superior colliculus, and has intimate reciprocal connections with a wide variety of cortical areas (Benevento & Fallon, 1975; Linke et al., 1999; Grieve et al., 2000; Kaas & Lyon, 2007). These neuroanatomical studies suggest that the pulvinar forms a subcortical visual route to the cortex that bypasses the striate cortex (Pessoa & Adolphs, 2010). Indeed, human subjects and monkeys with lesions in the striate cortex (V1) display a wide range of residual visual functions in the blind area (i.e. blindsight; Stoerig & Cowey, 1997). Monkeys with striate cortex lesions can discriminate spatial localization (Solomon et al., 1981), luminous flux (Pasik & Pasik, 1973), colors and figures (Schilder et al., 1972). Human subjects with V1 lesions can also respond differentially to spatial localization of stationary and moving stimuli (Perenin & Jeannerod, 1975; Blythe et al., 1987), motion direction (Barbur et al., 1980; Perenin, 1991), line orientation (Weiskrantz, 1987), wavelength (Morland et al., 1999) and form (Perenin & Rossetti, 1996). Consistent with these findings, some pulvinar neurons have retinotopically specific receptive fields and respond to moving stimuli with various directions, while the activity of other pulvinar neurons is modulated by spatial attention (Robinson, 1993). These pulvinar neurons might send visual information directly to the middle temporal area, accounting for some residual visual functions, especially spatial functions (Berman & Wurtz, 2010, 2011).

The pulvinar also projects to other subcortical areas such as the amygdala and striatum (Day-Brown et al., 2010; Pessoa & Adolphs, 2010; Tamietto & de Gelder, 2010). These subcortical routes might be involved in rapid processing of emotional stimuli (Tamietto & de Gelder, 2010). Indeed, patients with V1 lesions can differentially respond to gender and expressions in photos of humans (Morris et al., 2001), as well as line drawings of airplanes during fear conditioning (Hamm et al., 2003). In contrast, pulvinar damage impairs rapid processing of visual threat in humans (Ward et al., 2005). A recent neurophysiological study reported that monkey pulvinar neurons differentially respond to various emotional expressions in photos of humans (Maior et al., 2010). This subcortical pathway, comprising the superior colliculus, pulvinar and amygdala, is also implicated in rapid processing of facial information, including gaze direction (Johnson, 2005). Newborn babies with an immature cortical system preferentially orient toward faces with direct gaze and schematic face-like patterns (Johnson et al., 1991). Although this suggests pulvinar involvement in processing of facial and face-like stimuli, previous neurophysiological studies used only moving dots, gratings or simple patches. Consequently, evidence that pulvinar neurons process facial stimuli has been lacking. In the present study, we investigated neuronal responses to these stimuli in the monkey pulvinar.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Subjects

Two adult (one female and one male) macaque monkeys (Macaca fuscata), weighing 7.2–9.5 kg, were used in this experiment. Each monkey was individually housed with food available ad libitum. The monkeys were deprived of water and received juice as a reward during training and recording sessions. Supplemental water and vegetables were given after each day's session. To assess the monkeys' health, their weight was routinely monitored. The monkeys were treated in strict compliance with the United States Public Health Service Policy on Human Care and Use of Laboratory Animals, the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and the Guidelines for the Care and Use of Laboratory Animals of the University of Toyama. The study had been approved by the Committee for Animal Experiments and Ethics at the University of Toyama.

The monkey sat in a monkey chair 68 cm away from the center of a 19-inch computer display for behavioral tasks during the training and recording sessions in a shielded room. The CRT monitor was set so that its center was on the same horizontal plane as the monkey's eyes. The monkey chair was equipped with a responding button, which was positioned so that the monkey could easily manipulate it. An infrared charge-coupled device camera for eye-movement monitoring was firmly attached to the chair by a steel rod. During training and recording sessions, the monkey's eye position was monitored with 33-ms time resolution by an eye-monitoring system (Matsuda, 1996). The juice reward was accessible to the monkey through a small spout controlled by an electromagnetic valve. A PsyScope system (Carnegie Mellon University, Pittsburgh, PA, USA) controlled the electromagnetic valve and sound signal, as well as the timing of outputs to the CRT monitor.

Visual stimuli

Figure 1A shows the stimulus set, consisting of photos of human faces, used in the present study. These photos have been previously reported to activate monkey amygdalar neurons (Tazumi et al., 2010). The facial photos, obtained using five human models, consisted of three head orientations: straight ahead (frontal face); 30° to the right (profile face); and 30° to the left (profile face). The frontal faces consisted of three gaze directions (directed toward, and averted to the left or right of the monkey), and the profile faces comprised two gaze directions (directed toward, and averted to the right and left of the monkey). The facial stimuli were 256 digitized color-scale images. Stimuli were presented on a black background of 0.7 cd/m2 with their centers at the center of the display. The luminance of each stimulus was determined by measuring luminance of the circular area (radius, 6.35 cm) including each stimulus inside the circle by means of a luminance meter (BM-7A; Topcon, Tokyo). The luminance of these stimuli ranged from 1.36 to 3.66 cd/m2 [luminous intensity (total luminance) ranged from 16.4 to 44.2 mcd].

image

Figure 1. Visual stimulus set used in the present study. (A) Thirty-five facial photos of five different models, including two females (W1, W2) and three males (M1, M2, M3). The stimulus set for each model consisted of seven faces with different face orientations and gaze directions: (1) frontal view with direct gaze; (2) frontal view with gaze to the right; (3) frontal view with gaze to the left; (4) left profile view with direct gaze; (5) left profile view with indirect gaze; (6) right profile view with direct gaze; (7) right profile view with indirect gaze. (B) Fourteen artificial schematics, including three cartoon faces (C1–3) with three gaze directions, four face-like patterns (J1–4), three eye-like patterns with three gaze directions (E1–3), and four simple geometric patterns consisting of a circle, cross, square and star (S1–4).

Download figure to PowerPoint

We did not use facial stimuli with profiles rotated by 30° to the right and gaze direction averted to the right, or profiles rotated by 30° to the left and gaze direction averted to the left. In these facial stimuli, it is difficult to detect the dark iris; only the white sclera could be seen. In monkey faces, the iris can always be recognized as it occupies the major part of the visible eye. Therefore, this type of human facial stimuli appears to be unusual for monkeys. In addition, the iris can be recognized in all of the frontal faces, regardless of gaze direction, whereas in these particular profiles the iris cannot be recognized. The lack of the iris produces a qualitative difference among the facial stimuli. For these reasons, we avoided profiles without a visible iris.

Figure 1B shows line drawings of faces with three gaze directions (cartoon faces), eye-like patterns and face-like patterns (J1–4) that newborn babies orient toward (Johnson et al., 1991). The luminance of the white and black areas inside these illustrations was 36.5 and 0.7 cd/m2, respectively (total luminance of the cartoon faces, eye-like patterns and face-like patterns were 38.7, 188.6 and 179.3 mcd, respectively). In addition, as control stimuli, four simple geometric patterns (circle, cross, square and star) were used. Luminance of the white areas inside the simple geometric patterns was 36.5 cd/m2 (total luminance of the circle, cross, square and star were 151.6, 96.0, 188.1 and 61.0 mcd, respectively). The cartoon faces, eye-like patterns and face-like patterns comprised 256 digitized RGB images; the four simple geometric patterns comprised 256 digitized images. These stimuli were displayed on a CRT monitor with a resolution of 640 × 480 pixels, and the size of the stimulus area was 5–7 × 5–7°. Some of the pulvinar neurons were further tested with scrambled images of the stimuli that elicited the strongest responses. Scrambled images were formed by cutting the original image into 289–441 pieces and randomly reassembling the fragments.

Behavioral tasks

The monkeys were trained to perform a sequential delayed non-matching-to-sample (DNMS) task that requires discrimination of faces, face-like schematics and simple patterns (Fig. 1). The task was initiated by a buzzer tone; then, a fixation cross appeared on the center of the display. When the monkeys fixated on the cross for 1.5 s, a sample stimulus was presented for 500 ms (sample phase). The control phase was defined as the period of 100 ms before the sample phase. When facial photos were used as sample stimuli, gaze directions of the stimuli were either directed to or averted from the monkey. Then, after an interval of 1.5 s, the same stimulus appeared again for 500 ms, between one and four times (selected randomly for each trial). Finally, a new stimulus with different gaze direction was presented (target phase). When the target appeared, the monkey was required to press a button within 2 s to receive a juice reward (0.8 mL). When the monkey failed to respond correctly during the target phase or press the button before the target phase, the trials were aborted and a 620-Hz buzzer tone was presented. The inter-trial intervals were 15–25 s (Fig. 2).

image

Figure 2. Stimulus sequence in the delayed non-matching-to-sample task task, in which stimuli were sequentially presented with a delay between them.

Download figure to PowerPoint

In the DNMS task, the monkey compared a pair of stimuli in each trial (i.e. sample and target stimuli). Stimulus pairs consisted of the same category of stimuli; only pairs of facial stimuli and pairs of geometric patterns were used (i.e. facial stimuli were not paired with geometric patterns). In the facial pairs, averted gazes were always paired with directed gazes; stimulus pairs of gazes averted to the left and the right were not used. Furthermore, the facial stimuli presented in the target phase were the same as in the comparison phase, apart from gaze direction (i.e. same model and same head orientation); thus, the monkeys were required to detect a difference in gaze direction (directed vs. averted gaze). For the geometric patterns (Fig. 1B), only stimuli within the same category (cartoon faces, face-like patterns, eye-like patterns and simple geometric patterns) were paired. Thus, a total of 72 stimulus pairs (for each of the five models – frontal faces, four pairs; profile faces, four pairs; cartoon faces, four pairs; face-like patterns, 12 pairs; eye-like patterns, four pairs; simple geometric patterns, 12 pairs) were used. These procedures facilitated the monkeys in learning that a shift in gaze direction was an important clue for solving the task.

Training and surgery

The monkeys were trained in the DNMS task for 3 h/day, 5 days/week. The monkeys required about 11 months of training to reach a 97% correct-response rate. After completion of this training period, a head-restraining device (a U-shaped plate made of epoxy resin) was attached to the skull under aseptic conditions (Nishijo et al., 1988a,b; Tazumi et al., 2010). The subject was anesthetized with a combination of medetomidine hydrochloride (0.5 mg/kg, i.m.) and ketamine hydrochloride (5 mg/kg, i.m.). The plate was anchored with dental acrylic to titanium bolts inserted in the skull. We also implanted a reference pin, the location of which was based on the zero coordinates defined in the stereotaxic atlas of the brain of Macaca fuscata individuals (Kusama & Mabuchi, 1970). During the surgery, heart and respiratory functions and rectal temperature were monitored (LifeScope 14; Nihon Kohden, Tokyo, Japan). A blanket heater was used to keep body temperature at 36 ± 0.5 °C. Antibiotics were administered topically and systemically for 1 week after the surgery to prevent infection. Two weeks after surgery, the monkey was retrained while the head was painlessly fixed to the stereotaxic apparatus by using the head-restraining device. The performance criterion (> 85%) was again attained within 10 days.

Stereotaxic localization of the pulvinar for recording and histology

Before recording from the pulvinar in each hemisphere, a marker consisting of a tungsten wire (diameter – 500 μm) was inserted near the target area under anesthesia, and three-dimensional magnetic resonance imaging scans of the monkey head were performed. The 3D pictures of the monkey brain with the marker were reconstructed by computer rendering. The 3D stereotaxic coordinates of the target area were determined in reference to the marker in the 3D reconstructed brain (Asahi et al., 2003, 2006).

After the last recording session, several small marking lesions were created in the pulvinar by passing 20–30 μA of anodal current for 30 s through an electrode placed stereotaxically. Subsequently, the monkeys were deeply anesthetized with an overdose of sodium pentobarbital (60 mg/kg, i.m.) and perfused transcardially with 0.9% saline followed by 10% buffered formalin. The brains were removed from the skulls and cut into 50-μm sections containing the pulvinar. Sections were stained with Cresyl violet. The sites of electrical lesions were determined microscopically. The location of each recording site was then calculated by comparing the stereotaxic coordinates of recording sites with those of lesions, and were plotted on the actual tissue sections. Locations of visually responsive neurons in the two monkeys were compared on the basis of the shapes of the pulvinar nuclei, and were re-plotted on the serial sections of the pulvinar of one monkey, from 8 mm (AP8) to 5 mm anterior (AP5) to the interaural line.

Electrophysiological procedures and data acquisition

After the monkeys relearned the DNMS task at a > 85% correct ratio, we commenced recording neuronal activity. Neuronal activity was recorded from each hemisphere in both subjects. A glass-insulated tungsten microelectrode (0.8–1.5 MΩ at 1 kHz) was stereotaxically inserted into the pulvinar vertically to the orbitomeatal plane in a stepwise fashion by a pulse motor-driven manipulator (SM-21; Narishige, Tokyo, Japan). Only neuronal activities with a signal-to-noise ratio > 3 : 1 were recorded. The analog signals of neuronal activities, triggers for visual stimuli, juice reward, button pressing and the XY coordinates of eye position were digitized at a 40-kHz sampling rate and stored in a computer via a multichannel acquisition processor (MAP; Plexon, Dallas, TX, USA) system. This information was also recorded on a data recorder (RT-145T; TEAC, Tokyo). The digitized neuronal activities were isolated into single units by their waveform components using the Offline Sorter program (Plexon). Superimposed waveforms of the isolated units were drawn to assess variability throughout the recording sessions and transferred to the NeuroExplorer program (Nex Technologies, MA, USA) for further analysis. If the monkey exhibited signs of fatigue, such as closing the eyes for several seconds or moving the eyes or hands slowly, the experimental session was immediately terminated. In most cases, the unit recording experiment was terminated within 2–3 h. After responses to the 49 visual stimuli were recorded, the scrambled images were then presented to the monkeys if single unit activity was still observed.

Analysis of the basic characteristics of pulvinar neurons

Spike sorting was performed with the offline sorter program for cluster analysis (Off-line sorter, Plexon). Each cluster was checked manually to ensure that the cluster boundaries were well separated and the waveform shapes were consistent with the action potentials. For each isolated cluster, an autocorrelogram was constructed and only units with refractory periods > 1.2 ms were used for further analyses. Finally, superimposed waveforms of the isolated units were drawn to check the consistency of the waveforms. Figure 3A and B shows examples of superimposed waveforms of a pulvinar neuron and its autocorrelogram, respectively. This autocorrelogram indicates that the refractory period of the neuron was 2–3 ms throughout the recording sessions, which suggests that these spikes were recorded from a single neuron.

image

Figure 3. Identification of pulvinar neurons and analysis of pulvinar neuronal responses. (A) An example of superimposed traces of a pulvinar neuron. (B) Autocorrelograms of the neurons indicated in (A). Bin width = 1 ms. Ordinates indicate probability, where bin counts were divided by the number of spikes in the spike train. (C) A peri-event histogram of the neuron indicated in (A), showing responses to a visual stimulus. Bin width = 50 ms. The dashed line indicates onset of stimulus presentation, and the horizontal bar indicates the duration of the stimulus (500 ms). (D) Magnitudes of the pulvinar neuronal responses indicated in (C). The stimulus duration was divided into 10 epochs (50 ms each). Response magnitudes (spikes/s) were defined as follows – mean firing rate in each epoch minus mean firing rate during the 100-ms period before stimulus onset.

Download figure to PowerPoint

We analysed single neuronal activity during 500 ms after (‘post’) the onset of stimulus presentation in the sample phase, but did not analyse single neuronal activity in the target phase. Only stimuli that were presented more than five times in the sample phase were analysed. The baseline firing rate was defined as the mean firing rate during the 100-ms ‘pre’ period. Significant excitatory or inhibitory responses to each stimulus were defined by a Wilcoxon signed rank (WSR) test (< 0.05 for statistical significance) of neuronal activity between the 100-ms pre and the 500-ms post periods. Furthermore, to investigate temporal changes in neuronal responses, the 500-ms post period was divided into ten 50-ms epochs. The mean neuronal firing rate was calculated for each of these epochs. Response magnitude was defined as follows – mean firing rate in each epoch minus the mean firing rate during the 100-ms pre period. Figure 3C and D shows a peri-event summed histogram of responses from the same neuron shown in Fig. 3A and B to a facial photo (Fig. 3C) and response magnitudes in the 10 epochs converted from this histogram (Fig. 3D).

For each neuron, response magnitudes during the visual stimulation period (for the whole 500-ms period and for each epoch) for all 49 visual stimuli were analysed by one-way anova (one factor with 49 levels; < 0.05). Neurons with a significant main effect were defined as differential neurons. For each facial model, one-way anova was also performed. Responses to three frontal faces with three gaze directions, and those to right and left profile faces with two gaze directions were compared by Tukey post hoc tests (< 0.05). Neurons with significantly different responses toward gaze directions were defined as gaze-differential neurons (Tukey post hoc tests, < 0.05). Neurons with significantly different responses toward face orientations were defined as face orientation-differential neurons (Tukey post hoc tests, < 0.05). For other stimulus categories (cartoon faces, eye-like patterns, face-like patterns and simple geometric patterns), one-way anovas were also performed within the same stimulus category. Neurons with a significant main effect were defined as cartoon face-differential, eye-like pattern-differential, face-like pattern-differential and simple geometric pattern-differential neurons, respectively.

Stimulus information conveyed by visually responsive neurons (bits/s) was computed as described in previous studies (Skaggs et al., 1993; Panzeri et al., 1996). These parameters were calculated as follows,

  • display math

where I is the information rate of the neuron in bits/s, i is the stimulus number (code) to identify the 49 visual stimuli, λi is the response (mean firing rate) to the stimulus i, λ is the total average response, and P(i) is the probability of the stimulus. The neuronal activity of visually responsive neurons conveys stimulus information and, consequently, visually responsive neurons display high values for this parameter. This parameter was compared by one-way repeated-measures anova at a significance level of < 0.05. The post hoc comparisons were performed using the Bonferroni-corrected method with a significance level of < 0.05.

We also analysed response latency to each visual stimulus. For each neuron, one peri-event histogram was constructed using the entire set of data for all trials and all stimuli. Neuronal response latency was defined as the interval from the onset of stimulus presentation to the time at which the neuronal firing rate exceeded the mean ± 2 SD of the baseline firing rate. Furthermore, for each neuron, individual peri-event histograms were constructed using data for each of the different stimulus categories. We compared the latencies to various stimulus categories to determine whether the characteristics of the specific visual stimuli could modulate the latencies of the pulvinar neurons. All data were expressed as mean ± SEM.

Multivariate analysis of visual responses by pulvinar neurons

Multidimensional scaling (MDS) is a method used to simplify the analysis of relationships that exist within a complex array of data. MDS constructs a geometric representation of data to show the degree of relationship between stimuli represented by the data matrix (Young, 1987). MDS has been used to examine taste relationships in the gustatory system (Nishijo & Norgren, 1990, 1991), face categorization in the inferotemporal cortex (Young & Yamane, 1992) and spatial discrimination in the septal nuclei (Nishijo et al., 1997) by using data matrices representing neural activity in response to the particular stimulus array (i.e. taste solutions, photos of faces and photos of locations, respectively). In the present study, the 49 visual stimuli were used to elicit neural activity in pulvinar neurons.

Data matrices of neural activity in a 68 × 49 array derived from the 68 visually responsive neurons were generated. Euclidean distances as dissimilarity between all possible pairs of two visual stimuli were calculated by using the visual responses of the 68 pulvinar neurons. Then, the mds program (proxscal procedure, spss statistical package, version 16) positioned the visual stimuli in the two-dimensional space with the distances between the stimuli representing the original relationships (i.e. Euclidean distances in the present study; Shepard, 1962; Kruskal, 1964).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Responses to the visual stimuli

Recordings were made from a total of 401 neurons from the pulvinar nuclei of two monkeys. One-hundred and sixty-five neurons responded to visual stimuli and, of these, 68 neurons were tested with all of the visual stimuli. The mean spontaneous firing rate was 12.15 ± 1.14 spikes/s (= 68; mean ± SEM) and the mean firing rate during stimulus presentation (500 ms) was 24.67 ± 2.50 spikes/s (= 68).

The pulvinar visually responsive neurons showed robust responses especially during the first 100 ms after stimulus onset. Figure 4 shows such an example of a pulvinar neuron that responded to various visual stimuli. The activity of the neuron increased sharply in response to the onset of the stimuli, then decreased rapidly, and then gradually increased again. This pattern of changes in neuronal activity formed two response phases – an early rapid response phase and a late gradual response phase. This neuron responded strongly to the face-like patterns (Fig. 4G), especially in the late phase.

image

Figure 4. An example of a pulvinar neuron that responded to the visual stimuli. (A–E) Responses to the facial photos. (F–I) Responses to the face cartoons (F), face-like patterns (G), eye-like patterns (H) and simple geometric patterns (I). Horizontal bars above the raster displays indicate the stimulus presentation period (500 ms). The vertical dotted line in each of the raster displays and histograms indicates the stimulus-onset point. Calibration at the right bottom of the figure – number of spikes per trial in each bin. Bin width = 50 ms.

Download figure to PowerPoint

Figure 5A shows response magnitudes of the neuron shown in Fig. 4 during stimulus presentation (500 ms) of all of the visual stimuli. There were significant differences in response magnitudes to the various visual stimuli (F48,563 = 5.821, < 0.001; differential neuron). All of the 68 neurons tested displayed differential responses to the various stimuli (one-way anova,< 0.05). Furthermore, the neuron responded differentially to gaze direction in M2 and W1 (dotted lines; Tukey test, < 0.05). In addition, there were significant differences in mean response magnitudes to the five stimulus categories (F4,607 = 31.36, < 0.001). Subsequent post hoc tests indicated that mean response magnitude to the face-like patterns was significantly greater than those to the stimuli in the other stimulus categories (Tukey tests, < 0.001).

image

Figure 5. Response magnitudes during stimulus presentation (500 ms) to the visual stimuli. (A) Comparison of response magnitudes for the neuron shown in Fig. 4 for all visual stimuli. All of the 49 stimuli elicited significant excitatory responses (Wilcoxon signed rank, < 0.05). The neuron responded differentially to gaze directions of models M2 and W1 (dotted lines; Tukey tests, < 0.05). In addition, mean response magnitudes to the face-like patterns were significantly greater than those to the stimuli in the other categories (Tukey tests after one-way anova,< 0.001). For each model, responses to the stimuli with different face orientation and gaze directions are similarly aligned from the left, as for model M1. (B) Comparison of response magnitudes of the 68 visually responsive neurons to the face-related stimuli (facial photos, cartoon faces, eye-like patterns and face-like patterns) and non-face stimuli (simple geometric patterns). The mean response magnitude to the face-related stimuli was larger than that to the non-face patterns (F1,3330 = 5.76, *< 0.05). (C) Comparison of response magnitudes of the 68 visually responsive neurons to the five stimulus categories. The face-like patterns and eye-like patterns elicited stronger responses than the simple geometric patterns (Tukey tests, ***< 0.001 and **< 0.01, after one-way anova).

Download figure to PowerPoint

The overall mean responses indicated that the pulvinar neurons responded stronger to the face-related stimuli (facial photos, cartoon faces, eye-like patterns and face-like patterns) than the non-face stimuli (simple geometric patterns). Figure 5B illustrates the mean response magnitudes of the 68 visually responsive neurons during stimulus presentation (500 ms) to the face-related and non-face stimuli. The mean response magnitude of the 68 visually responsive neurons to the face-related stimuli was significantly larger than that to the non-face stimuli (F1,3330 = 5.76, < 0.05). Figure 5C shows the mean response magnitudes of the 68 visually responsive neurons to the five stimulus categories. There were significant differences in response magnitudes to the five stimulus categories (F4,3327 = 26.67, < 0.001). The face-like and eye-like patterns elicited stronger responses than the simple geometric patterns (Tukey tests, < 0.001 and 0.01, respectively). These results indicate that the pulvinar neurons responded well to face-related stimuli.

Response patterns of the individual pulvinar neurons

Of these 68 visually responsive neurons, 23 neurons responded differentially to gaze direction in the frontal or profile faces of at least one of the facial models (gaze-differential), and 29 responded differentially to face orientation (face orientation-differential). Differential responses were exhibited by nine neurons to gaze direction of cartoon faces (cartoon face-differential), and by four neurons to gaze direction of eye-like patterns (eye-like pattern-differential). Five and eight neurons responded differentially to face-like patterns (J1–4; face-like pattern-differential) and simple geometric patterns (simple geometric pattern-differential), respectively. Ratios of the gaze-differential and face orientation-differential neurons (23/68 = 33.8% and 29/68 = 42.6%, respectively) were significantly higher than those of the cartoon face-differential (9/68 = 13.2%), eye-like pattern-differential (4/68 = 5.9%), face-like pattern-differential (5/68 = 7.4%) and simple geometric pattern-differential neurons (8/68 = 11.8%; Fisher's exact probability test, all < 0.01).

A previous study by our group demonstrated that the mean response magnitudes toward facial photos with direct gaze were significantly larger than those to facial photos with averted gaze in the monkey amygdala (Tazumi et al., 2010). We analysed the pulvinar responses in the same manner. However, the difference in response magnitudes to these two different gaze directions was not statistically significant in the pulvinar nuclei (paired t-test, > 0.05).

Figure 6 shows the results of a cluster analysis of the 68 neurons based on the response magnitudes to the 49 stimuli during the 500-ms period after stimulus onset; although typical clusters were not observed, groups of neurons with similar response trends were identified. Units 1–34 (Cluster J) comprised neurons that responded best or second best to one of the face-like patterns, except for five neurons (units 15, 19, 28, 33 and 34). Units 35–39 (Cluster C/E) consisted of neurons that responded best or second best to one of the cartoon faces and eye-like patterns. Units 40–54 (Cluster W) responded best or second best to one of the facial photos of the female models, except for two neurons (units 43 and 46). Clusters Ma, Mb and Mc comprised neurons that responded best or second best to facial photos of the different male models. Cluster S consisted of neurons that responded best or second best to the simple geometric patterns. This classification of visually responsive neurons based on cluster analysis does not necessarily imply the existence of distinct clusters of neurons, as the response patterns of these neurons continuously changed, both between and within clusters. Nonetheless, these results demonstrate that the activity of pulvinar neurons is modulated according to the stimulus category.

image

Figure 6. Dendrogram of the 68 pulvinar neurons tested with the 49 stimuli resulting from hierarchical cluster analyses. Units 1–34 (Cluster J) comprise neurons that responded best or second best to one of the face-like patterns. Units 35–39 (Cluster C/E) include neurons that responded best or second best to one of the cartoon faces and eye-like patterns. Units 40–54 (W) responded best or second best to one of the facial photos of the female models. Clusters Ma, Mb and Mc contain neurons that responded best or second best to facial photos of the three different male models, respectively. Cluster S includes neurons that responded best or second best to the simple geometric patterns. The unit number and response category of each neuron based on best and second best stimuli are listed on the left. For abbreviations of the best stimuli (Fig. 1B). The right and left stimuli in the abbreviations indicate the best and second best stimuli, respectively. Abscissa, cluster similarity between neurons or clusters.

Download figure to PowerPoint

The above response patterns of the pulvinar neurons indicate that the pulvinar neurons were also more responsive to the face-related stimuli than the non-face stimuli (simple geometric patterns). Among the five categories of the visual stimuli, ratios of the pulvinar neurons that responded best to the face-like patterns and facial photos (27/68 = 39.7% and 22/68 = 32.3%, respectively) were significantly higher than those of the pulvinar neurons that responded best to the eye-like patterns, cartoon faces and simple geometric patterns (11/68 = 16.2%, 3/68 = 4.4% and 5/68 = 7.4%, respectively; Fisher's exact probability test, all < 0.05). These results indicate that the pulvinar neurons were more responsive to the face-like patterns and facial photos than the eye-like patterns, cartoon faces and simple geometric patterns.

Responses to scrambled stimuli

To analyse whether the visual responses were dependent on a coherent pattern of visual stimuli, we compared responses to optimal stimuli with responses to scrambled images of those stimuli. Figure 7A and B shows examples of two pulvinar neurons tested with scrambled images. The neuron shown in Fig. 7A responded strongly to the face-like patterns (Aa–Ac) but less to the scrambled image (Ad), while the neuron shown in Fig. 7B responded strongly to the human frontal faces (Ba–Bc) but less to the scrambled image (Bd). Figure 7C shows the effects of scrambling of the stimuli. Scrambling significantly reduced responses to the facial photos (paired t-test, < 0.05) and face-like patterns (paired t-test, < 0.001). These results indicate that the visual responses of the pulvinar neurons were dependent on coherent visual patterns present in the stimuli.

image

Figure 7. Examples of two pulvinar neurons tested with the scrambled images. (A) The neuron responded strongly to the face-like patterns (Aa–Ac), but less to the scrambled image (Ad). (B) The neuron responded strongly to the human frontal faces (Ba–Bc), but less to the scrambled image (Bd). Horizontal bars above the raster displays indicate the stimulus presentation period (500 ms). A vertical dotted line in each of the raster displays and histograms indicates the stimulus-onset point. Calibration at the right bottom of the figure – number of spikes per trial in each bin. Bin width = 50 ms. (C) Effects of scrambling of the stimuli. Scrambling significantly reduced the responses to the facial photos (paired t-test, *< 0.05) and face-like patterns (paired t-test, ***< 0.001).

Download figure to PowerPoint

Response latencies of the pulvinar neurons

Response latencies were analysed for all of the 165 visually responsive neurons. Figure 8A shows the mean response latencies of the pulvinar neurons to various visual stimuli. The distribution of the latencies formed two peaks – a short latency group (30–120 ms) and a long latency group (170–500 ms). The mean latency of the short latency group was 63.38 ± 1.89 ms. There was no significant difference in mean latencies between the lateral and medial pulvinar (62.03 ± 2.34 ms vs. 65.61 ± 3.56 ms, t-test, > 0.05).

image

Figure 8. Distribution of the response latencies of the 165 visually responsive neurons in the pulvinar. (A) The distribution of the mean response latencies. In each neuron, mean response latency was estimated using responses to all visual stimuli. The distribution of the latencies formed two peaks – a short latency group (30–120 ms) and a long latency group (170–500 ms). (B) Comparison of response latencies among the stimulus categories in the short latency group. The mean response latencies to the face-like patterns (J1–4) were significantly shorter than those to the other stimulus categories (Tukey tests after one-way anova, ***< 0.001). (C) Comparison of response latencies between the face orientations. The mean response latencies to the frontal faces were significantly shorter than those to the profile faces (paired t-test, **< 0.01).

Download figure to PowerPoint

To investigate how configuration of visual stimuli modulates the response latencies, we analysed the response latency to each category of visual stimuli (Fig. 8B). In the short latency group, there were significant differences in response latencies to the various stimulus categories (one-way anova; F4,205 = 11.446, < 0.001). Multiple post hoc comparisons indicated that the mean response latencies to the face-like patterns (J1–4) were very short (50.12 ± 1.58 ms), shorter than those to the other categories (Tukey test, < 0.001). There was a significant difference in response latencies to the various facial photos as well (Fig. 8C). The mean response latency to the frontal faces (62.67 ± 1.49 ms) was significantly shorter than that to the profile faces (66.00 ± 1.73 ms; paired t-test, < 0.01).

Temporal changes in response magnitudes to the visual stimuli

Figure 9 shows response magnitudes in four different epochs of the same neuron shown in Fig. 4. In epoch 1, during the first 50-ms period (Fig. 9A), this neuron showed strong responses to the face-like patterns; three of the face-like patterns (J1, 2, 4) elicited stronger responses than stimuli from the other categories, and the remaining face-like pattern (J3) elicited stronger responses than stimuli from the other categories, except for seven stimuli (Tukey test after one-way anova,< 0.05). Furthermore, the most face-like patterns (J1) elicited stronger responses than the other face-like patterns (J2, 3, 4; Tukey tests after one-way anova,< 0.05). In epoch 2, during the second 50-ms period, from 50 to 100 ms after stimulus onset (Fig. 9B), all of the visual stimuli elicited significant excitatory responses (WSR test, < 0.05). Furthermore, the neuron responded differentially to gaze direction in M2, M3 and W1 (dotted lines; Tukey tests, < 0.05) and to face orientations in W2 (solid lines; Tukey test, < 0.05). In epoch 3, during the third 50-ms period, from 100 to 150 ms after stimulus onset (Fig. 9C), only one cartoon face elicited inhibitory responses, while most other stimuli elicited excitatory responses (WSR test, < 0.05). Furthermore, the neuron responded differentially to gaze direction in W1 and W2 (dotted lines; Tukey tests, < 0.05). In epoch 10, during the last 50-ms period, from 450 to 500 ms after stimulus onset (Fig. 9D), the face-like patterns elicited stronger responses than some other stimuli. These findings suggest that neuronal responses to visual stimuli were different in different epochs.

image

Figure 9. Temporal changes of response patterns of the neuron shown in Fig. 4 across the different epochs. (A) Comparison of response magnitudes to the 49 visual stimuli in epoch 1. This neuron responded strongly and specifically to the face-like patterns (J1–4). Upward columns, excitatory responses; downward columns, inhibitory responses; NS, non-significant responses (Wilcoxon signed rank, > 0.05). *Significant difference between the face-like patterns (Tukey tests after one-way anova,< 0.05); #significant difference from the Johnson-3 except the stimuli with this mark. (B) Comparison of response magnitudes to the 49 visual stimuli in epoch 2. The neuron responded to all of the different stimulus categories and responded differentially to gaze directions of models M2, M3 and W1 (dotted lines; Tukey tests, < 0.05), and face orientations of model W2 (solid lines; Tukey tests, < 0.05). (C) Comparison of response magnitudes toward the 49 visual stimuli in epoch 3. The neuron responded differentially to gaze directions of models W1 and W2 (dotted lines; Tukey tests, < 0.05). (D) Comparison of response magnitudes toward the 49 visual stimuli in epoch 10. The other descriptions are similar to those for Fig. 5.

Download figure to PowerPoint

Figure 10 shows the mean response magnitudes of the 68 visually responsive neurons in four different epochs. The data again revealed similar trends. In epoch 1, the face-like patterns elicited stronger responses than the other visual stimuli (Tukey test after one-way anova,< 0.01). In epoch 2, response magnitudes to all visual stimuli increased; the mean response magnitude to each stimulus was significantly larger than in epoch 1 (paired t-test, < 0.05). These results suggest that pulvinar neurons are more sensitive to visual stimuli in epoch 2. These changes in responsiveness were not uniform across the various visual stimuli at the single neuron level; the neurons displayed differential responses to these stimuli.

image

Figure 10. Temporal changes in mean response patterns of the 68 visually responsive neurons across the different epochs. (A) Comparison of response magnitudes toward the 49 visual stimuli in epoch 1. In epoch 1, the face-like patterns (J1–4) elicited stronger responses than the other visual stimuli (Tukey tests after one-way anova,< 0.001). (B) Comparison of response magnitudes toward the 49 visual stimuli in epoch 2. In epoch 2, response magnitudes to all visual stimuli increased compared with epoch 1 (paired t-test, < 0.05). (C) Comparison of response magnitudes for the 49 visual stimuli in epoch 3. (D) Comparison of response magnitudes for the 49 visual stimuli in epoch 10. The other descriptions are similar to those for Fig. 5.

Download figure to PowerPoint

Figure 11A shows the number of differential neurons (one-way anova,< 0.05) in each epoch. The number of differential neurons was significantly higher in epoch 2 than in epoch 1 (Fisher's exact probability test, < 0.001). Furthermore, the amount of information conveyed by the 68 pulvinar neurons was significantly higher in epochs 2 and 3 than in epoch 1 (Tukey test after one-way anova,< 0.05; Fig. 11B). These results suggest that pulvinar neurons send more information on visual stimuli to upstream visual areas in epoch 2 than in epoch 1.

image

Figure 11. Changes in the number of differential neurons (A) and amount of information conveyed by the pulvinar neurons (B) across the 10 epochs. (A) Number of differential neurons in each epoch. ***Significant difference from epoch 1 (Fisher's exact probability test, < 0.001). Number in each histogram indicates number of differential neurons. (B) Amount of information conveyed by the pulvinar neurons in each epoch. **,*Significant difference from epoch 1 (Bonferroni-corrected t-test after one-way repeated-measures anova,< 0.01, 0.05).

Download figure to PowerPoint

Multivariate analysis of pulvinar neuronal responses

The above analyses suggest that pulvinar neurons specifically encode face-like patterns in epoch 1 and supplementary information in epoch 2. The data sets of the response magnitudes recorded from the 68 pulvinar neurons in epochs 1 and 2 were subjected to MDS analysis (Figs 12 and 13). After calculating stress values and squared correlations (R2) for up to four dimensions, we chose a two-dimensional space (Bieber & Smith, 1986). For the two-dimensional solutions, the R2 values for epochs 1 and 2 were 0.957 and 0.737, respectively. In epoch 1 (Fig. 12), one cluster without face-like patterns (J1–4) was recognized. In this large cluster, the stimuli in the four stimulus categories (facial photos, cartoon faces, eye-like patterns and simple geometric patterns) were intermingled. The face-like patterns formed a separate small group. These data also suggest that, in the first 50-ms period, pulvinar neurons specifically process visual information of face-like patterns. In epoch 2 (Fig. 13), the five clusters corresponding to the five stimulus categories (i.e. facial photos, cartoon faces, face-like patterns, eye-like patterns and simple geometric patterns) were recognized. These results are consistent with the changes in information amount in epoch 2 and indicated that, in the second 50-ms period after stimulus onset, the pulvinar neurons processed more information on the visual stimuli.

image

Figure 12. Distributions of the 49 visual stimuli in a two-dimensional space resulting from multidimensional scaling using responses of the 68 neurons to these stimuli in epoch 1. Two groups were recognized – one large group, including the facial photos, cartoon faces, eye-like patterns and simple geometric patterns; and one small group, including only the face-like patterns (J1–4). C1–3, three cartoon faces; J1–4, four face-like patterns; E1–3, three eye-like patterns; S1–4, simple geometric patterns; W1–2, female models; M1–3, male models. The numbers from 1 to 7 after the model numbers and colons indicate face orientations and gaze directions indicated in Fig. 1A. Inset – neurons surrounded by dashes are shown, with magnified coordinates.

Download figure to PowerPoint

image

Figure 13. Distributions of the 49 visual stimuli in a two-dimensional space resulting from multidimensional scaling using the responses of the 68 neurons to these stimuli in epoch 2. Five groups corresponding to the five stimulus categories (i.e. facial photos, cartoon faces, face-like patterns, eye-like patterns and simple geometric patterns) are recognized. The other descriptions are similar to those for Fig. 12.

Download figure to PowerPoint

Locations of the pulvinar neurons

We recorded neuronal activity from various subnuclei of the pulvinar, which mainly included the lateral pulvinar, medial pulvinar and inferior pulvinar. Histological data indicated that all of the visually responsive neurons were located within the pulvinar. Distributions of the visually responsive (open circles) and non-responsive (dots) neurons are illustrated in Fig. 14. Most of the responsive neurons were distributed in the lateral and medial pulvinar.

image

Figure 14. Recording sites of the 401 pulvinar neurons. (A–F) The neurons are plotted on coronal sections of the left pulvinar at different A-P levels. The number in the left upper corner of each section indicates distance (mm) anteriorly from the interaural line. The horizontal axis indicates the distance (mm) from the midline; the vertical axis indicates the distance (mm) from the interaural line. Open circles, visually responsive neurons; dots, non-responsive neurons. Because some symbols are superimposed, the total number of the symbols is smaller than the total number of pulvinar neurons (= 401). CC, corpus callosum; Cd, caudate nucleus; eml, external medullary lamina; GC, substantia grisea centralis; GMpc, nucleus genicularus medialis pars parvocellularis; IC, inferior colliculus; MD, medial dorsal thalamus; OPr, nucleus olivaris pretectalis; PI, inferior pulvinar; PL, lateral pulvinar; PM, medial pulvinar; R, reticular thalamus; SC, superior colliculus; SG, nucleus supragenicularus; W, white matter.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

General characteristics of the pulvinar neurons

The visually responsive neurons were located mainly in the dorsal lateral pulvinar and ventral part of the medial pulvinar in the present study. In contrast with the retinotopically organized region in the ventral lateral pulvinar (Benevento & Port, 1995; Kaas & Lyon, 2007), the medial pulvinar, anterior dorsal and caudal ventral parts of the lateral pulvinar are non-retinotopic regions, where neurons respond differentially to some patterns and/or colors, and have large, bilateral and binocular receptive fields, including the fovea (Benevento & Miller, 1981; Felsten et al., 1983; Benevento & Port, 1995). The caudal ventral part of the lateral pulvinar receives inputs from superficial layers of the superior colliculus (Harting et al., 1980) and prestriate cortices (Benevento & Davis, 1977), and projects to the inferotemporal cortex (Benevento & Rezak, 1976). The recording sites in this study correspond to these non-retinotopic regions.

The pulvinar neurons displayed a broad distribution of response latencies, ranging from 30 to 500 ms. The mean latency of the short latency group was 63.38 ± 1.89 ms, which was comparable to previous studies (Felsten et al., 1983; Benevento & Port, 1995). Because the mean latency in V1, which projects to the pulvinar, is 66 ± 10.7 ms (Schmolesky et al., 1998), some pulvinar neurons with short latencies, especially those with latencies < 60 ms, might receive inputs from the superior colliculus or directly from the retina. The remaining pulvinar neurons in the short latency group with latencies > 66 ms might receive inputs from the various visual cortices. In comparison, the pulvinar neurons in the long latency group might receive inputs from some other structures, such as the temporal association cortices, prefrontal cortex or the amygdala, which project to the pulvinar (Shipp, 2003). In addition, response latencies to frontal faces were significantly lower than those to profile faces. This suggests that the subcortical visual pathway might be tuned better to frontal faces than to profile faces.

Pulvinar responses to forms

Total luminance of the face-like patterns was smaller than those of the square and eye-like stimuli, and luminance of the white areas of the face-like patterns was the same as that of the simple geometric patterns, indicating that specific early responses to the face-like patterns are not due to differences in total luminance or luminance. Furthermore, scrambling of the images greatly reduced the pulvinar responses in the present study. This is the first evidence that pulvinar neuronal responses are dependent on coherent facial patterns. The results also indicate that selective responses to some visual stimuli are attributable to factors other than luminance differences. These findings are consistent with previous studies on face neurons in the prefrontal cortex (Ó Scalaidhe et al., 1999), inferotemporal cortex (Desimone et al., 1984) and superior temporal cortex (Bruce et al., 1981). However, in contrast to these cortical facial areas and the amygdala (Tazumi et al., 2010), the responses of the pulvinar neurons were not specific to faces; pulvinar neurons also responded to other visual stimuli, such as simple geometric patterns, in the present study. Furthermore, although there was no significant difference in mean response magnitude toward faces with direct and averted gazes, many individual pulvinar neurons differentially responded to gaze directions (i.e. gaze-differential neurons). Neurophysiological and human imaging studies have shown that the amygdala responds stronger to faces with direct gaze (Kawashima et al., 1999; Wicker et al., 2003; Sato et al., 2004; Tazumi et al., 2010). These findings suggest that the pulvinar sends information on gaze direction to higher upstream brain areas in the visual pathway, such as the amygdala (Tazumi et al., 2010) or superior temporal cortex (Perrett et al., 1985; De Souza et al., 2005), which receive afferents from the pulvinar (Yeterian & Pandya, 1991; Shipp, 2003).

It has been reported that Type II neurons in the superficial layer of the superior colliculus have relatively large receptive fields, including the parafoveal area, and respond differentially and strongly to complex forms but poorly to conventional stimuli (stationary or moving white and dark spots and slits; Rizzolatti et al., 1980). The lateral pulvinar neurons' differential responses to forms might receive inputs from these Type II neurons. The medial pulvinar receives inputs from the deep layer of the superior colliculus (Benevento & Fallon, 1975; Linke et al., 1999; Grieve et al., 2000), which receives input from the superficial layer of the superior colliculus (Isa, 2002; Doubell et al., 2003), and has reciprocal connections with various association cortices and the amygdala (Grieve et al., 2000; Shipp, 2003). These anatomical connections and pulvinar neuronal responses to coherent patterns provide the anatomical and neurophysiological bases of the subcortical visual pathway for specific form detection.

Role of the subcortical visual pathway

Two hypotheses regarding the role of the subcortical visual pathway have been proposed. One hypothesis states that the subcortical visual pathway (superior colliculus–pulvinar–amygdala) might convey fast and coarse information (Johnson, 2005; Day-Brown et al., 2010; Tamietto & de Gelder, 2010). Several studies provide evidence for the existence of a subcortical visual pathway for fast and coarse information processing of faces. Human neurophysiological studies using Magnetoencephalographies reported short latency responses (30–60 ms), for which sources were presumed to be located in the pulvinar (Braeutigam et al., 2001). In the present study, the responses to the face-like patterns were very selective in epoch 1 (first 50-ms period). These short latencies in the present study are comparable to those in the human study (Braeutigam et al., 2001). Recent studies indicate that holistic face perception is largely supported by low spatial frequencies and suggested that holistic processing precedes the analysis of local features during face perception (Goffaux & Rossion, 2006), and face contours (similar to the face-like patterns in the present study) shortened response latencies to faces in the human occipito-temporal regions (Shibata et al., 2002). Low spatial frequency information is important for face recognition in newborn babies with relatively immature visual cortical areas (Johnson, 2005; de Heering et al., 2008). The face-like patterns used in the present study are the same as those used in the experiment using newborn babies, in which the newborn babies preferentially oriented toward such stimuli (Johnson et al., 1991). These face-like patterns are equivalent to low spatial frequency components of faces (Johnson et al., 1991). The present study shows that the pulvinar neurons were most sensitive to face-like patterns in epoch 1. Consistently, low-frequency faces specifically activate the subcortical visual pathway, including the superior colliculus, pulvinar and amygdala (Vuilleumier et al., 2003). Furthermore, residual visual ability was tuned to low spatial frequency in a patient with blindsight due to lesions in the visual cortical areas (Sahraie et al., 2002). This fast activation of the pulvinar might be due to direct inputs from the superior colliculus, contributing to the ability of newborns to orient toward faces. The present study provides neurophysiological evidence of pulvinar involvement in fast and coarse facial information processing.

The second hypothesis proposes that interactive activity based on reciprocal connections between the subcortical and cortical areas is important for stimulus recognition and attention (Bullier, 2001; Pessoa & Adolphs, 2010). These cortico-pulvino-cortical circuits might be involved in coordinating and amplifying signals, and improving signal-to-noise ratios (Shipp, 2003; Pessoa & Adolphs, 2010), as well as modulating interactions between oscillatory processes in different cortical areas, which contributes to visual attention (Serences & Yantis, 2006; Saalmann & Kastner, 2009). Our results here indicate that pulvinar neurons detect face-like patterns in epoch 1, while they categorize the visual stimuli into one of the five stimulus categories in epoch 2. Furthermore, the amount of stimulus information conveyed by the pulvinar neurons and the number of stimulus-differential neurons was higher in epoch 2 than in epoch 1. These results indicate that pulvinar neurons become more sensitive to other categories of stimuli after epoch 1 (i.e. epoch 2 or later), during which cortical neurons also become active (for response latencies of cortical neurons, see a review by Lamme & Roelfsema, 2000). These findings suggest that pulvinar responsiveness to a variety of stimuli in epoch 2 might be due to reciprocal connections with cortical areas with similar response latencies. Consistent with this, a neuropsychological study of human patients with pulvinar lesions suggests that the pulvinar is involved in enhancing stimulus saliency (Snow et al., 2009), which might contribute to neural computation in an early stage of stimulus categorization (Meeren et al., 2008).

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Our results provide direct neurophysiological evidence that pulvinar neurons respond to face-like patterns with short latencies, which seems to be consistent with the view that the pulvinar nuclei comprise a subcortical pathway that rapidly processes coarse facial information. Following the initial recognition of the facial stimulus, the population activity of the pulvinar neurons participates in classifying the facial pattern, with a concomitant increase in the amount of information processed. These results suggest two functions of the pulvinar – providing the circuitry to detect specific stimuli, such as faces, and the processing of stimulus saliency and visual attention based on connections with cortical areas.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

The authors thank Dr Robert H. Wurtz (NIH) for valuable comments on this manuscript. This research was supported in part by the JSPS Asian Core Program, the Ministry of Education, Science, Sports and Culture, a Grant-in-Aid for Scientific Research (A) (22240051), and the National Bio-Resource Project (NBRP) ‘Japanese Monkeys’ of the MEXT, Japan.

Abbreviations
DNMS

delayed non-matching-to-sample

MDS

multidimensional scaling

WSR

Wilcoxon signed rank

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References