Table 1.
The demographic data.
Figure 1.
Experiment 1: Detection of biological motion in noise.
A. Each trial consisted of two successive 1 sec presentations of PL animations (separated by a 0.5 sec blank period), with one interval containing biological motion in noise and the other interval containing scrambled motion in noise; the scrambled motion on each trial was always derived from the biological motion presented on that trial. The left panel shows one frame depicting biological motion in the first interval (black dots indicate biological motion) and the right panel a frame of scrambled motion in the second interval. In the actual experiment all dots appeared as black against a white background. Noise dots had the same local motion trajectories as those of the biological or scrambled motion on that trial. The set of biological motion sequences totaled 24 distinct activities: 5 walking (stairway walking, climbing, crossing a small object, and 2 plain walking with different viewing angle), 4 jumping (standing jump, leaping, rope-jumping, and high-jumping), 4 kicking (toward front, side, and 2 soccer kicking), 2 running (plain and turning around), 6 throwing (3 overhead and 3 under-throwing), and 3 crouching. B: Results of the biological motion detection task. Mean detection thresholds for the two groups are shown, together with error bars indicating ±1 standard error of the mean (SE).
Table 2.
The demographic data.
Figure 2.
Experiment 2: Discrimination of perturbation of biological motion.
A: A series of parametrically perturbed motion sequences was created from 10 different PL animations each depicting a different human activity. These ten different PL animations comprised 2 portraying jumping (standing jump, rope-jumping), 3 kicking (toward front, toward side, and soccer kicking), 3 throwing (tossing, bowling, overhead throwing), 1 crouching for high jump, and 1 backward walking. In this example, black dots indicate the dots forming a single frame of a normal biological PL sequence and gray dots illustrate the corresponding frame of spatially scrambled version of this sequence. For example, dot A' indicates a new location of dot A when the motion is 100% spatially scrambled. The position denoted as (a) corresponds to an intermediate position that divides the distance between A and A' in the ratio of 15∶85. Therefore, when the position ‘(a)’s are taken from all the other pairs of biological-scrambled dots, a sequence containing 15% perturbed biological motion is generated. In the same way, (b),(c), and (d) represent the dot positions of 30%, 45%, and 60% perturbed motion. B: Single frame exemplars of the four discrimination conditions. Over trials, these pairs of animations portraying differing degrees of perturbation were presented in random order, and following each trial the participant indicated which one (left or right) was closer to unperturbed human motion. The % values below each figure refer to the percent of spatial perturbation. C: Performance (accuracy of discrimination) on the task in the schizophrenia group (filled symbols) and the healthy control group (open symbols). Error bars indicate ±1 standard error of the mean (SE). Chance performance on this 2AFC task corresponds to 50% correct.
Table 3.
The number of subject who performed above chance accuracy of binomial distribution in each perturbation level.
Figure 3.
Experiment 2: Performance for each distinct activity.
Each activity animation was quantified in terms of the total motion energy in that animation (defined by the total excursion of dots over space and time during the 1 sec presentation). The motion energy of each animation is speicified by the y-axis on the left-hand side of the graph, and the animations are ordered from most to least motion energy along the x-axis. (1 = standing jump; 2 = kicking side; 3 = crouching jump; 4 = soccer kicking; 5 = kicking front; 6 = backward walking; 7 = bowling; 8 = rope jumping; 9 = tossing; 10 = overhead throwing). The histogram bars show average discrimination performance associated with each activity pooled across all perturbation pairings except the most extreme perturbations where performance on the task was impossible. Filled bars are for schizophrenia patients and open bars for healthy controls. There is no correlation between performance and total motion energy.
Table 4.
The demographic data.
Figure 4.
Experiment 3: Localization of regions of interest.
A. Stimuli and procedures used to localize area STSp. Shown on the left are examples of PL biological motion and scrambled motion, and on the right is shown schematically the block-designed runs for STSp localization. B. Stimuli and procedures used to localize area MT. Dots moving radially inward and outward and static dots were presented in block-designed runs for MT localization. C. Inflated whole-brain images (both hemispheres for one patient and for one healthy control) showing regions of interest identified using the localizers described above.
Figure 5.
Experiment 3: Event-related portion of brain imaging study.
A. The schematics of successive animation frames shown on the left are examples of the three categories of PL animations presented in an event-related fMRI design shown on the right. B. Mean(SE) d' on the biological motion task performed during the event-related functional scan. C. Hit and false alarm rates associated with the d' values shown in panel B.
Table 5.
Talairach coordinates of the defined ROIs.
Figure 6.
Experiment 3: Brain imaging results from STSp.
A: Average time-series associated with each of three signal detection categories (hit, correct rejection, false alarms) in controls (left) and patients (right). Time value 1 denotes stimulus onset (TR = 2 sec). B: Each histogram plots, for patients and healthy controls, the peak BOLD signal levels (1 SE) associated with each of the three signal detection categories. C. Same as panel B, with data removed for the two schizophrenia patients for whom STSp localization was based on anatomy, not differences in activations on the STSp localizer.
Figure 7.
Experiment 3: Brain imaging results from MT.
Same format as in Figure 6.