Gaze fixation and the neural circuitry of face processing in autism

Abstract

Diminished gaze fixation is one of the core features of autism and has been proposed to be associated with abnormalities in the neural circuitry of affect. We tested this hypothesis in two separate studies using eye tracking while measuring functional brain activity during facial discrimination tasks in individuals with autism and in typically developing individuals. Activation in the fusiform gyrus and amygdala was strongly and positively correlated with the time spent fixating the eyes in the autistic group in both studies, suggesting that diminished gaze fixation may account for the fusiform hypoactivation to faces commonly reported in autism. In addition, variation in eye fixation within autistic individuals was strongly and positively associated with amygdala activation across both studies, suggesting a heightened emotional response associated with gaze fixation in autism.

Figure 1

Study designs and exemplar stimuli. (a) Study I: 2 (Emotion) × 2 (Orientation) repeated-measures design. Twenty-four emotional faces and 16 neutral faces were used, half with eyes and face oriented straight ahead, and half with face and eyes averted 45 degrees (equal toward the right and left). (b) Study II: examples of matched familiar versus unfamiliar photographs. Photographs of participants’ family and friends were matched for gender, age, facial expression and orientation as closely as possible to photographs taken by other study participants.

Figure 2

Average fixation durations. (a) Study I, average duration of fixation on the mouth and eye region and face in general broken down by group. (b) Study II, average duration of fixation on the mouth and eye region and face in general broken down by group. Error bars index the s.e.m.

Figure 3

Brain clusters with significant group differences in brain activation across all faces in Study I. (a) Right fusiform gyrus: x = 32, y = −57, z = −3; 2,142 voxels. (b) Left fusiform gyrus: x = −39, y = −57, z = −6; 508 voxels. (c) Right occipital gyrus: x = 14, y = −85, z = 13; 904 voxels. (d) Left occipital gyrus: x = −7, y = −91, z = 6; 670 voxels. (e) Right middle frontal: x = 42, y = 7, z = 32; 463 voxels. (f) Left orbitofrontal gyrus: x = −24, y = 36, z = −10; 119 voxels. (g) Left amygdala: x = −22, y = −8, z = −18; 122 voxels. All images are presented in radiological convention such that the right hemisphere is at left of each coronal image. Clusters are color-coded based on the control-minus-autism t-statistic values (positive values indicate control values greater than autism). The clusters are also superimposed on an averaged echo-planar blood oxygenation level–dependent imaging (EPI BOLD) signal illustrating adequate signal coverage for each cluster. Averaged MR time series are presented below each cluster for 14 s post-stimulus onset.

Figure 4

Brain clusters with significant group differences in brain activation across all faces in Study II. (a) Right fusiform gyrus: x = 48, y = −53, z = −11; 125 voxels. (b) Left anterior medial fusiform gyrus: x = −25, y = −65, z = −1; 322 voxels. (c) Left posterior lateral fusiform gyrus: x = −33, y = −73, z = −6; 292 voxels. (d) Right occipital gyrus: x = 6, y = −78, z = −1; 304 voxels. (e) Right amygdala: x = 21, y = −2, z = −20; 31 voxels. The clusters are color-coded based on the control-minus-autism t-statistic values (positive values indicate control greater than autism). The clusters are also superimposed on an averaged EPI BOLD signal illustrating adequate signal coverage for each cluster. Averaged MR time series are presented below each cluster for 14 s post-stimulus onset.

Figure 5

Brain clusters associated with significant Group × Familiarity interactions in brain activation in Study II. (a) Right fusiform gyrus: x = 28, y = −76, z = −10; 226 voxels. (b) Right occipital gyrus: x = 18, y = −78, z = −7; 318 voxels. Clusters are color-coded based on the Group × Familiar F-statistic values. The clusters are also superimposed on an averaged EPI BOLD signal illustrating adequate signal coverage for each cluster. Averaged MR time series are presented below each cluster for 14 s post-stimulus onset. Data for familiar stimuli are in the top graphs and unfamiliar stimuli in the bottom graphs.

Figure 6

Brain activation clusters associated with average eye fixation time for the autistic and control groups, Study I. (a) Left amygdala: x = −19, y = −6, z = −13; 106 voxels. (b) Right anterior fusiform gyrus: x = 28, y = −50, z = −11; 76 voxels. The clusters are also superimposed on an averaged EPI BOLD signal illustrating adequate signal coverage for each cluster. Scatter plots showing the relationship between brain activation and average eye fixation are given for each group (autistics, top graphs; controls, bottom graphs) below each cluster. The regression line and 90% confidence bands are superimposed on each scatter plot.

Figure 7

Brain activation clusters associated with average eye fixation time for the autistic and control groups, Study II. (a) Right amygdala: x = 11, y = −7, z = −20; 63 voxels. (b) Right anterior fusiform gyrus: x = 27, y = −57, z = −6; 113 voxels. The clusters are also superimposed on an averaged EPI BOLD signal illustrating adequate signal coverage for each cluster. Scatter plots showing the relationship between brain activation and average eye fixation are given for each group (autistics, top graphs; controls, bottom graphs) below each cluster. The regression line and 90% confidence bands are superimposed on each scatter plot.

Figure 8

Clusters in the left amygdala associated with group differences in activation as a function of amount of eye fixation within subjects. (a) Study I: right amygdala: x = 18, y = −3, z = −12, 110 voxels; control M = −1.31, s.d. = 2.18, autism M = 2.68, s.d. = 3.18, group t_1,16 = −3.11, P = 0.006. (b) Study II: right amygdala: x = 15, y = −8, z = −18, 46 voxels; control M = −1.64, s.d. = 5.21, M = 4.46, s.d. = 5.94, group t_1,28 = −2.96, P = 0.006. The clusters are also superimposed on an averaged EPI BOLD signal illustrating adequate signal coverage for each cluster. Bar graphs showing the β-weight derived for each group using average eye fixation to predict brain activation are given below each cluster. Error bars index the s.e.m.