By Kuan-Ming Chen Nov. 14th, 2005.

Treisman, A. (1998). Feature binding, attention and object perception. Philosophical Transactions of the Royal Society of London: Series B, Biological Sciences, 353, 1295-1306. By Kuan-Ming Chen Nov. 14th, 2005.

Outline • The Binding Problem • A Role for Spatial Attention? • Evidence from Illusory Conjunctions • Visual Search and Binding • Evidence from Parietal Lesions • Binding in Feature Search • Implicit Processing of Conjunctions • Conclusions

The Binding Problem • How do we specify what goes with what and where? • We simply are not aware that there is a problem to be solved.

Two Important Facts • Binding problem is more acute under that • The world is a crowded place, full of objects. • Single neurons respond across areas that would certainly hold several objects in crowded displays.

Anatomical & Physiological Evidence • Specialized brain areas code different aspects of the visual scene. • Dorsal vs. ventral pathway: motion, space vs. color, shape (Ungerleider & Mishkin, 1982) • Achromatopsia: color (Meadows, 1974; Damasio et al., 1980) • Akinetopsia: motion (Zihl et al., 1983; Zeki, 1991) • Focal activity shifts to different brain areas as subjects respond to different aspects of the same displays.

Which Mechanisms for Binding? • Directly code conjunctions of features? • IT responds to relatively complex combinations of features (Tanaka, 1993) • However… • Only one object was shown at a time in these studies. • IT could be coding the output of the binding process. • Too few neurons to code individually the combinatorial explosion of arbitrary conjunctions.

Which Mechanisms for Binding? • Synchronized neural activity • Units that fire together would signal the same object (Gray et al., 1989; Fries et al., 2001; Singer & Gray, 1995). • However, false bindings happen when different objects share the same features.

A Role for Spatial Attention? • Posner’s (1980) paradigm of visual attention • Visual search paradigm

A Role for Spatial Attention? Facilitation • Posner’s (1980) paradigm of visual attention • Attention window

A Role for Spatial Attention? • Posner’s (1980) paradigm of visual attention • Visual search paradigm

A Role for Spatial Attention? • Visual search paradigm • RT of conjunction search ∝ display size • Attention window is centered on each object in turn. (Treisman & Gelade, 1980)

How Spatially Selective Attention Plays a Role in Binding? • We code one object at a time, selected on the basis of its location. • By temporarily excluding stimuli from other locations, we can simply bind whatever properties are currently attended.

A map of locations • It registers the locations of regions without the features. • A set of feature maps • A flag signaling the presence of the feature.

Feature search • By checking the flag for the presence of features, no attention is needed (e.g., red, green, horizontal etc.). • Conjunction search • An attention window moves within the location map and selects from the feature maps whatever features are currently linked to the attended location (e.g., horizontal red).

T O TorX (color intrusion) TX (letter intrusion) (illusory conjunction) Evidence from Illusory Conjunctions • R.M. • Patient with severe problems in binding features (Balint syndrome) • He made illusory conjunction error more than 35% of trials.

How about normal subjects? • Illusory conjunctions in normal subjects • Example

To report the digits first, then all the features of a stimulus cued by a bar marker 6 3 Sample display Quatre

Digits? Stimulus?

How about normal subjects? • Illusory conjunctions in normal subjects • Conjunction errors recombining the shape, size, format and colour in 18% of trials, compared with only 6% intrusion errors.

O X H O X How about normal subjects? • Are there two identical items in this display? • NO!! • But, even same-different matching task reveals binding errors without a load on memory.

How spatial attention prevents errors • To focus attention on target item • To cue the relevant item 150 ms before the display, and allowed subjects to ignore the digits (to focus attention on target) • Binding errors disappeared! • Spatial attention plays a role in the binding process.

The number of instances of features (b) (a) • An illusory red bar in (a) and (b) were made at 1.5:1 (not 3:1), although there are three times as many red objects in (a).

Therefore… • Features are separately coded, otherwise they cannot recombine. • The binding problem is a real one. • Focused attention is involved in solving it. • Attention is not required for the simple detection of separate features.

Visual Search and Binding • Feature search • No effect of the number of items. • Conjunction search • Attention is focused on each item in turn to find conjunction targets, RTs increase linearly with the number of items. (Treisman & Gelade, 1980)

Modified model • In some conjunction search a feature-based strategy is used. • E.g., search for O among Xs and Os • Search for O among O (color) while inhibiting X • Search for O among X (shape) while inhibiting O

Evidence from Parietal Lesions • Hypothesis • Parietal lobes are involved in spatial attention, and master map of location depends on parietal function. • Evidences • TMS to the right parietal lobe slowed conjunction search but left feature search unaffected (Ashbridge et al., 1997). • R.M.→ strokes of bilateral parietal lobes.

Four Difficulties by R.M. • A risk of illusory conjunctions whenever more than one object is present. • Conjunction search is abnormally difficult; however, feature search is no difficulty. • Only one object is visible at a time (individuating objects depends on binding them to separate locations) • Difficulties in conscious access to spatial information such as pointing, reaching etc.

Binding in Feature Search • Search for feature targets could be difficult if they closely resemble the distractors. • What counts as a feature for the visual system must be determined empirically. • Stimuli differing only slightly along a single dimension would not activate separate feature detectors.

Asymmetries in search difficulty A curved lines could be coded as basically straight, plus some additional activity in detectors for curvature. Purple target require binding of activity in blue and red detectors that share the same location (Treisman, 1991).

Implicit Processing of Conjunctions • Three new directions • Measured by implicit indices? • Manifested in memory? • The binding of actions to perceived events?

Can implicit binding be revealed? • To read the word • R.M.’s RTs were slower when the word was in an inconsistent location. • Implicit information about locations is coded. Up Down Down Up Inconsistent Consistent

Can implicit binding be revealed? • R.M. named the letter on the side that experimenter had tapped on the shoulder at chance level. • But to name the first letter he saw, he reported significantly more from the tapped side. • Some implicit representation of space remains.

Implicit processing in normal subjects Prime Negative priming To match green shape with the right one Probe However, if size of probe is changed, facilitation is found.

Facilitation Prime Prime Probe Facilitation when we recombined two unattended halves. To match the shapes in the two outside positions

Implicit processing in normal subjects • Negative priming implies • The subjects formed and retained a memory trace even when it was the unattended one; • An action tag was bound to the memory. • The inhibition persisted 200 trials later. • However, when we recombined them or changed the size of probe, the result was facilitation.

Conclusions • Attention is needed to bind features together. • Without attention, the only information recorded is the presence of separate parts and properties. • Binding problem may be bound up with the nature of consciousness.

Posner’s Detection task(To find a black square) Quatre

By Kuan-Ming Chen Nov. 14th, 2005.