[back]

Attention modulation of cell firing in visually responsive neurons in the human medial temporal lobe
Leila Reddy, Patrick Wilken, Christof Koch

Does attention modulate cell firing in visually responsive neurons from hippocampus, para-hippocampal gyrus, and the amygdala in the human brain?

To investigate this question, we intend to record from single units in medial temporal lobe structures while human subjects (patients undergoing epilepsy monitoring with intra-cranial electrodes prior to epilepsy surgery) view pictures previously shown to elicit category selective visual responses (Kreiman et al., 2000). Attention will be manipulated during the experiment and changes in neural responses will be examined when attention is directed towards versus away from the pictures.

The first principal problem in sensor based motion planning is the find-goal problem. In this problem, the robot seeks to use its on-board sensors to find a collision free path from its current configuration to a goal configuration. In the first variation of the find goal problem, which we term the absolute find-goal problem, the absolute coordinates of the goal configuration are assumed to be known. A second variation on this problem is described below.

To manipulate attention during human subject recordings, we plan to utilize a visual-tracking paradigm similar to that employed by Pylyshyn and Storm(1988). Each trial begins by briefly presenting a set of objects (i.e., disks) and highlighting a subset of them, which is tracked to the end of the trial while subjects fixate on a central fixation point. A period of approximately 5 seconds follows while the objects move around on the computer monitor. At the end of the trial, one object is highlighted and the subject indicates whether this object was in the original target set by a button press. The speed and density of the display is adjusted to ensure that removal of attention from the tracked objects will result in a miss. While participants perform the object-tracking task, a set of images of faces is presented in the background for 1s each and with an inter-stimulus interval of 500 ms. A schematic illustration of this design is shown in Figure 1. In order to focus attention on the picture set, participants are required to perform a gender discrimination task concurrently. As will be discussed below, preliminary results show that performance on the gender discrimination task is modulated by attentional load.

Figure 1. A schematic illustration of one trial in the Object Tracking experiment. A set of objects is shown on the computer monitor and a subset of them is highlighted for 500 ms. The objects then move around randomly on the monitor for 2000 ms and observers are required to track the targets. During this task, a set of pictures will be randomly presented in the background for 1000 ms each and observers will perform a gender discrimination task. At the end of the trial, one object will be highlighted and observers must indicate whether this object was in the original target set.

There is literature showing that varying the number of items tracked alters the attentional load demands of the task (Jovicich et al., 2001). Assuming that attention is a limited resource (e.g., Braun, 1999), it follows that increasing the attentional load in the tracking task leads to less attention being allocated to performance on the picture set. Thus, if it can be shown that in a high attentional load condition, the tracking task consumes attention which cannot then be allocated to the pictures, any modulation in the firing rate between a high-load and low-load condition can be attributed to changes in the amount of visual attention allocated to the pictures.

With the current design, in order to infer that neuronal modulation is due to attention, it is necessary to show that attention is removed from the background pictures in the high-load condition, and is available in the low-load condition. If the ball-tracking task consumes attention that is required for the picture task in the high-load condition, interference will be observed when participants perform the dual task. If attention is available to the picture task in the low-load condition, less interference will be observed in the dual task condition. Currently, experiments are being initiated in a non-clinical population to demonstrate that performance in the gender discrimination task is indeed compromised when subjects simultaneously perform the tracking task in the high-load condition.

Face-gender discrimination modulated by attentional load
Recent work in our laboratory has shown that rapid natural-scene categorization (animal vs. non-animal) can be done with little or no attentional cost (Li et al.). For this study, we investigated the attentional cost associated with a task requiring fine discrimination of stimuli sharing similar features: discrimination of face gender. Subjects were required to perform a face gender discrimination task (face database obtained from MPI, Germany) either alone or concurrently with an attentionally demanding task (5-letter T/L discrimination). Both letters and images were masked following presentation ( SOA 106ms for faces, variable for letters, ranging from 173ms to 200ms). Face gender discrimination performance was significantly lower in the dual-task condition compared to the single-task condition, demonstrating that performance in this task is modulated by attentional load(Figure2). On the other hand, it is interesting to notice that percentage correct in the dual-task condition was still highly above chance level, implying that in the near absence of attention, observers can distinguish the gender of a face to some extent. This property is surprising from a computational point of view since subjects cannot distinguish between the letters T and L under precisely the same conditions, as shown recently in our group.

Figure 2. Normalized average performance for 3 subjects with 2 different image sets. Each point is a subject's average dual task performance normalized according to his/her single task performance. Normalized values are obtained by a linear scaling which maps the average single task performance to 100%, leaving chance at 50%. A significant decrease in performance is observed for each subject in the dual task condition compared to the single task condition. Image sets were randomly chosen from the face database used. Blue circles: image set 1, Red circles: image set 2.

References

Brain Areas Specific for Attentional Load in a Motion Tracking Task. Jovicich, J., Peters, R. J., Koch, C., Braun, J., Chang, L., & Ernst, T. (2001). (submitted)

Category-specific visual responses of single neurons in the human medial temporal lobe. Kreiman, G., Koch, C., & Fried, I. (2000). Nature Neuroscience, 3(9), 946-953.

Attentional capacity is undifferentiated: Concurrent discrimination of form, color, and motion. Lee, D.K., Koch, C., & Braun, J. (1999). Perception and Psychophysics, 61(7), 1241-1255.

Rapid Natural Scene Categorization Without Attention. Li, F., VanRullen, R., Koch, C., & Perona, P. (2001). (in preparation)

Tracking multiple independent targets: evidence for a parallel tracking mechanism. Pylyshyn, Z. W., & Storm, R. W. (1988). Spatial Vision, 3(3), 179-197.