Face Social Preference (FSP)

Fig. 5 FSP and interest area placement and sizing.

Overview

The FSP task is an eye-tracking task to test social orienting and decreased social motivation with regards to gaze. Each trial in this gaze-contingent task begins with a 100*100px ball animation. Once gaze is within the ball’s interest area (IA), two still images of faces (or a in one condition a toy) are displayed oriented away from screen center (as shown in Fig. 5 ) The images and the resulting video are sized to 533*533px``and are displayed centrally within the left or right half of the screen (an eccentricity from center of ``480px).

Conditions

Away vs Towards (AT)

Videos of caregivers facing away vs towards

Invariant vs Variant (IV)

Videos of caregivers saying “hello”/”good job” (variant) vs smiles (invariant).

Face vs Toy

Shown videos of a caregiver’s faces vs videos of toys.

Interest Areas

There are three interest areas regions for the experiment. The 250px ellipse surrounding the ball and the left and right rectangular 600*600px regions around each image/video. The name of each interest area changes based on behavior, e.g. if the right video is a ‘variant’ stimuli and the right variant side is triggered to play, the IA for that side is labelled “Variant_Video_IA” and the non-video side is labeled Invariant_Still_IA. This labelling is to make clear which IA side and condition was showing video, and which was a still image. This labelling could aid analysis and dwell time measures.

Gaze Triggers

Each trial begins with the ball animation which rotates until gaze is detected to the ball IA region by an invisible boundary (IB) trigger. This trigger has a minimum duration set to 100ms. Gaze must be continuously within the ball IA region for 100ms for the trigger to fire. Once this trigger fires, the stimuli as still images are shown (Fig. 5). For trials that do not have a fixed choice video side, there are IB triggers that will fire if gaze is held within the still image IA for 100ms (minimum duration). The triggered side will then show the corresponding video (after a prespecified delay).

Classifiers

There is no built-in classifier for this task as the interest area analysis and eye movement or behavior of interest to the videos will need to be deduced/coded after data collection. There is, however, clear labelling of which side was triggered by gaze. There is also an initial look RT measure which provides an RT value for the first sample of gaze detected to the still image that triggers the video (after 100ms).

Animations

The only animation is the 100*100px ball which rotates clockwise at 1Hz.

Event Messaging

Each trial begins with the message BALL_ANIMATION_ONSET and is followed by GAZE_TRIGGER_TO_BALL when gaze triggers the IB, or after 2500ms without a gaze trigger the message TIMEOUT_FALSE_START is written. Following the gaze-trigger the message STILL_IMAGE_ONSET signifies the onset of the still images. Gaze triggering the left video writes the message GAZE_TRIGGER_LEFT, and the right trigger GAZE_TRIGGER_RIGHT. If the video presentation side is fixed by the data source, the message FIXED_SIDE is written. Once the video begins, there is a message for each frame of the video, e.g. ‘RIGHT VIDEO FRAME 1’ with a numeric value showing the currently displayed frame. This was included to allow a simpler frame level interest area values to be matched up if required. If not gaze triggers a stimulus within 5000ms TIMEOUT_NO_GAZE is written and the trial is recycled. The trial ends with the message DISPLAY_BLANK when the final blank screen is displayed.

Event Marker	Away-Towards vs Towards-Away		Face vs Toy		Invariant vs Variant		EGI DIN	EYELINK DIN
	Looked-Away	Looked-Towards	Face	Toy	Invariant	Variant
BALL_ANIMATION_ONSET	dafp						DIN3	4
GAZE_TRIGGER_TO_BALL	gafp
TIMEOUT_FALSE_START
STILL_IMAGE_ONSET	dsfp						DIN2	2
GAZE_TRIGGER_LEFT	glfp
	vlta	vlat	vlft	vltf	vliv	vlvi	DIN3	4
GAZE_TRIGGER_RIGHT	grfp
	vrta	vrat	vrft	vrtf	vriv	vrvi	DIN3
FIXED_SIDE	vlta / vrta	vlat / vrat	vlft/vrft	vltf/vrtf	vliv/vriv	vlvi/vrvi	DIN3
TIMEOUT_NO_GAZE
DISPLAY_BLANK	blfp, brfp

In addition to the event codes above, the following codes are also sent to the EGI Netstation acquisition system, but generally are not needed for offline analysis:

Description	EGI Event
Begin trial	bgn1
begin task	VBeg
Netstation misc event	TSYN
display start menu	dstr
display break menu	dbrk
display end screen	dend
display attention screen	attn
display blank screen	dbsp, dbfp
Gaze at left video	glfp
Gaze at right video	grfp
Display Blank Left	blfp
Display Blank Right	brfp

Understanding the EGI event codes

In the EGI event codes, the first letter indicates the type of event, and the remaining letters indicate the event. For example, vlta indicates display video left, looking towards to away. vlat stands for display video left, looking away to towards the participant.