The research at CNAS is generally based on the Theory of Visual Attention (TVA) first proposed by Bundesen (1990) and later developed into a Neural Theory of Visual Attention (NTVA; Bundesen, Habekost, & Kyllingsbæk, 2005). According to TVA, visual perception of the world around us may be thought of as a race between multiple perceptual categorizations of the objects in our visual field. The first elements to finish processing with respect to some categorization (winners of the race) are encoded into a limited visual short-term memory (VSTM) for consciousness and action. The total capacity for processing the categorizations is limited and distributed among the objects in proportion to their attentional weights. The amount of allocated processing capacity determines how fast a categorization can be encoded into VSTM. The processing rate of a particular perceptual categorization ("element x belongs to category i") is given by the Rate Equation:
where the processing speed (v value) of the categorization is given by a combination of three factors: the strength of the sensory evidence that element x has feature i, η(x, i), the perceptual bias related to category i, βi, and the relative weight of element x (the weight of element x, wx, divided by the sum of the weights of all other elements in the visual field S). The total visual processing speed, C, equals the sum of all v values:
C = Σ v
The cognitive parameters of the encoding process can be measured using a whole report task (Duncan et al., 1999; Habekost & Starrfelt, 2009), where single or multiple visual elements (typically letters) are flashed at varying exposure durations. The task is simply to identify as many of the elements as possible. Importantly, reports are unspeeded and accuracy is the only important measure. This implies that there is no significant motor component in the task: performance depends specifically on the efficiency of the visual system. The probability of encoding letter stimuli (measured by number of correct reports) develops systematically as a function of exposure duration. Analyzed by the TVA equations, the performance data yield estimates of three central parameters of attentional capacity: the perception threshold, t0, the visual processing speed, C, and the storage capacity of visual short-term memory, K. If distractor elements are added to the display (e.g., letters of another colour than the target objects; partial report) the parameter α can also be measured by comparing the accuracy reduction in this condition to trials without distractor elements (see Kyllingsbæk, 2006, for details of the estimation procedure). α is defined as the attentional weight of a distractor relative to a target object and can be taken as a measure of the distractibility of the participant.
Bundesen, Habekost, and Kyllingsbæk (2005) proposed an interpretation of TVA's Rate Equation at the level of single neurons in the visual system. According to this model, NTVA, the attentional weight of an object corresponds to the number of neurons that respond to its properties. This number can be varied by dynamic remapping of receptive fields (Moran & Desimone, 1985) such that signals from attended objects have a higher probability of being gated to a given neuron (and thus control its response). By a complementary mechanism, varying the perceptual bias for making certain categorizations (the β parameter of the Rate Equation) corresponds to up- or downscaling the activity in neurons specialized for making the categorization (see Treue & Martinez-Trujillo, 1999, for an example). In the NTVA paper we used this simple model to account in detail for sixteen central studies in the single-cell literature on visual attention, covering many of the central findings in the field (see also Bundesen & Habekost, 2008).
A central hypothesis of the research at CNAS is that arousal influences visual perception by modulating the β parameter in TVA's Rate Equation. NTVA links the β parameter directly to neural activation. Perceptual bias is a specific type of neural activation that favours particular categorizations. Arousal, on the other hand, is a non-specific type of neural activation. The new hypothesis is given exact form in the Arousal-Bias Equation, which we propose as an elaboration of TVA's Rate Equation:
βi = rvisual pi ui
rvisual represents the average responsiveness of neurons in the visual system. The value of this parameter is not linked to any particular visual categorization, but represents a general measure of the current excitability of the visual system. The perceptual bias is assumed to derive from two components: an expectation and a utility component. pi is a measure of the subjective prior probability (i.e., the expectation) of being presented with stimuli that have feature i in the next moment. ui is a measure of the subjective utility of detecting feature i (i.e., the expected benefit of making the categorization i). In contrast to the unspecific rvisual parameter, both pi and ui are related to particular categorizations. Further, in analogy to visual processing speed C (cf. Equation 2), we propose that the total visual arousal, Avisual, equals the sum of all visual β values:
Avisual = Σ β
Finally, we plan to investigate the hypothesis that the total visual arousal is given as a variable proportion of the brain's total arousal, Ageneral, depending on the relative attentional priority of the visual modality, Pvisual,
where S denotes the set of all sensory, motor, and cognitive representation modalities (e.g., auditory, somatosensory, olfactory, visual, motor, verbal etc.). Equation 5 implies that visual arousal can be up- or downscaled in accordance with the subject's prioritization of categorizations from the visual modality.