PRE-ATTENTIVE PROCESSING

Pre-attentive processing is defined as the initial process of automatic and rapid detection of salient objects in the environment without having to focus attention on them. It typically refers to the set of information captured by the human eye in a 'single glance', which falls within the range of 200-250 milliseconds. Various neurological studies state that features like hue, curvature, size, orientation, length, intensity, motion, and depth of field are immediately detected by the human eye during this short time. These features thus help humans to perform visual tasks such as target detection, boundary detection, and counting and estimation. Humans have evolved to discern patterns in the environment and use them to their advantage. ‘It represents a deep instinct, a drive, a need to impose order on the world so as to make it usable and survivable’ (Brigitte Jordan, 2013). In this paper, we will first investigate the neurological attributes that contribute to the detection of the pre-attentive features and later we will focus on the theories that help us understand the human behaviour of perceiving the surroundings as patterns. In the end, we will evaluate the design aspects of the YouTube website. YouTube is an American online social media platform mainly used for streaming and sharing videos.

​

Neurological attributes responsible for pre-attentive processing

Visual scenes comprise a lot of information, only a fraction of which is picked up by the eyes for further processing. The eyes make ‘rapid, ballistic movements’ (Purves D, Augustine GJ, Fitzpatrick D, et al., 2001) known as saccades to change the point of fixation by directing the fovea (area of highest visual acuity) onto new stimuli. The vision depends on the information that is picked during the fixation pauses between saccades (Min Zhao et al., 2012) which usually varies between 200-400 ms. The information once picked flows from the photoreceptors through the ganglion cells to the visual cortex.

​

Retina to Visual Cortex

     The retinal output is efficiently condensed so that the information-carrying axons can pass through the anatomical bottleneck caused by the optic nerve along the route from the eye to the visual cortex (Jonathan J. Nassi & Edward M. Callaway, 2009). The information is then projected to the lateral geniculate nucleus (LGN) located in the posterior part of the thalamus.  LGN is a six-layered structure that comprises neurons with concentric receptive fields (Chris I. Baker, 2012). The midget ganglion cells form the origin of the parvocellular pathway and convey the red/green color opponent signals to the upper four parvocellular (P) layers comprising of smaller neurons. Parasol ganglion cells form the origin of the magnocellular pathway and convey broadband, achromatic signals to the deepest two magnocellular (M) layers comprising of larger neurons. Finally, the bistratified ganglion cells form the koniocellular pathway that conveys the blue-on/ yellow-off color opponent color signals to the thin koniocellular layers intercalated between the six primary layers (Jonathan J. Nassi & Edward M. Callaway, 2009; Chris I. Baker, 2012). The smaller P cells of the LGN are involved in the analysis of form and color, whereas the larger M cells mostly participate in the detection of motion and luminance (Ehud Kalpan, 2003). Ultimately, the axons of all the LGN neurons terminate in different layers or sub-layers of the primary visual cortex through the occipital lobe.

Visual Cortex

     The visual cortex is the primary cortical region of the brain with the primary function to process visual information (Trevor Huff et al., 2021). ‘Most visual information from the LGN passes through V1 before being processed further in extrastriata visual cortex’ (Jonathan J. Nassi & Edward M. Callaway, 2009). Neurons in V1 are usually associated with the detection of ‘color, direction, and orientation’ (Chris I. Baker, 2012). V1 comprise areas that have colour-sensitive neurons and are known as blobs. The area between the blobs is referred as interblobs and are orientation sensitive (Stewart Shipp, 1995).  The receptive fields of the primary visual cortex (V1) are mainly classified as simple, complex and hypercomplex. The simple cells primarily respond to the orientation of lines and edges and the complex cells respond to the movements in a specific direction (Trevor Huff at al., 2021). The hypercomplex cells on the other hand respond to the orientation of corners and curves. Neurons with these patterns of response act as line detectors, motion detectors, and angle detectors respectively. (Trevor Huff at al., 2021) V2 receives its inputs from V1. Research shows that the neurons in this region respond to changes in color, complex patterns, spatial frequency, and orientation. The information from V2 further splits into the dorsal and ventral streams which are concerned with the processing of object recognition and spatial task / visual motor skills respectively.

​

Theories of pre-attentive processing

 

Feature Integration Theory

Treisman & Gelade proposed the Feature Integration Theory (FIT) of attention first in 1980 ( Jeremy M Wolfe, 2020) that suggest ‘features are registered early, automatically, and in parallel across the visual field, while objects are identified separately and only at a later stage, which requires focused attention’ (Terisman & Gelade, 1980). The theory further suggests that the ‘visual scene is initially coded along a number of separable dimensions, such as color, orientation, spatial frequency, brightness, direction of movement’ ( Terisman & Gelade, 1980). Healey & Enns explained Treisman’s feature integration model of early vision that states, each feature map registers the activities of only a specific visual feature, which are encoded parallelly. However, the feature maps do not provide any details about the location, spatial arrangement. Thus, focused attention is required to ‘glue’ the ‘initially separable features into unitary objects’(Terisman & Gelade, 1980). Healey & Enns further explained that the targets with the unique features can simply access the feature map to check if the activity is occurring and since feature maps are encoded parallelly, feature detection becomes instantaneous. However, conjunction targets can only be detected by searching serially through the maser map which requires focused attention. Thus, this model provides a generic hypothesis to explain the pre-attentive process.

​

Grouping: Pattern Perception Principles

Multiple studies have provided evidence that suggests perceptual grouping occurs pre attentively and does not require attention (Pedro R. Montoro et al., 2014). Gestalt psychologist in the early twentieth century identified a set of principles/factors of perceptual grouping (Joseph L Brooks, 2014). Gestalt principles focused on the grouping of elements based on proximity, similarity, closure, good continuation, common region, and connectedness (Irvin Rock & Stephen Palmer, 1990). Proximity is one of the most important grouping principles that can override competing visual cues such as similarity of color or shape (Aurora Harley, 2020). Various ‘phenomenological observation tells us that the strength of the proximity principle decreases with distance’ (Michael Kubovy et al., 1998). Aurora Harley also states that the elements that show identical visual traits such as color, shape or size are perceived to be related. Thus the ‘strength of the similarity principle decreases with dissimilarity’ (Michael Kubovy et al., 1998). Similarly, elements that either appear in a closed boundary, appear on a line/ curve, or seem connected by uniform visual traits are perceived to be related. Recent studies by Rosenthal & Humphreys and Wang, Weng, & He have reported evidence of pre-attentive processing of other perceptual attributes like ‘contour integration’ and ‘illusory contour completion’ respectively. Rosenthal and Humphreys also discovered that the integration of Gabor elements resulted in unconscious learning of global contours. These studies make imposing examples of the unconscious detection of the distinct elements to form patterns that are comparatively different from the sum of their parts (Pedro R. Montoro et al., 2014).

 

Guided Search Theory

The guided search theory by Wolfe hypnotized a construction of an activation map based on both bottom-up and top-down visual search (Christopher G. Healey & James T. Enns, 2012). ‘Activation map is the signal that will guide the deployment of attention. For each item in a display, the guiding activation is simply a weighted sum of the bottom-up activation and the activity in each channel (composed of the top-down activation) plus some noise’ (Jeremy M. Wolfe, 2007).’Bottom-up activation follows feature categorization. It measures how different an element is from its neighbors’ (Christopher G. Healey & James T. Enns, 2012). The more an item differs from its surroundings, the more it pops out and this happens because of local contrast (Jeremy M. Wolfe, 2007).