Funded Grants

Orienting attention to mental representations

Orienting attention to mental representations is a research proposal at the core of human cognitive neuroscience. The proposal has two nested goals. The broader objective is to tackle one of the fundamental questions in the field: What are the units of correspondence between cognitive function and brain function. The specific objective is to develop a new experimental paradigm to investigate the nature of mental representations and our ability to orient attention to their selective attributes.

Mapping cognitive functions onto brain functions

Human cognitive neuroscience is a long-awaited butterfly, just taking flight. People have scratched their heads for thousands of years trying to link what was inside mentally with what was inside physically.

Recent methods that measure brain activity non-invasively have boosted significantly the scientific investigation of the human mind. Positron emission tomography and functional magnetic resonance imaging detect changes in blood flow that follow increased metabolic demands; and reveal the brain regions that participate during a behavioral task. They provide maps of brain activity with high anatomical resolution at the brain-region scale, but with poor temporal resolution. Recordings of electrical or magnetic potentials associated with perceptual, cognitive or motor events are useful for testing the reactivity of the brain to experimental variables. Though poor in spatial resolution, recordings have real-time resolution and the sensitivity to capture brief changes in neural activity. Convergent findings from imaging and recording methods can build a dynamic view of the distributed neural systems that achieve cognitive functions.

The picture is macroscopic. Methods in human cognitive neuroscience provide little or no information about the computations or mechanisms that take place at the levels of molecules, single cells, or local cellular micro-circuitry within brain regions. Nevertheless, they have the power to address one of the most fundamental questions about brain and cognition - the nature of their mapping.

Much of our vision of cognition is still influenced by folk psychology and introspection. We divide cognition into common-sense domains of perception, attention, memory, language, etc. Scientific experimentation often respects these domain boundaries. Research fields within domains develop their own vocabulary and theories. There is an implicit assumption that the mapping of cognition onto neural function also obeys these domain boundaries. If so, there would be distinct brain circuits responsible for functions within perception, attention, memory, language, etc. This view is a modern version of phrenology - with less elaborate categories of cognitive functions, and neural circuits substituting for brain regions.

But what if common sense leads us astray from the true fabric of cognition? (For example, physics is leading us to a remarkably weird view of the nature of physical reality.) Instead of being bound by cognitive domains, brain areas may contribute particular functions that are useful for multiple cognitive domains. Some evidence for the sharing of brain areas across domains already exists. For example, associative visual areas in inferior temporal cortex have been shown to participate in perception [1], attention [2], and long-term associative memory [3]. Mapping between cognitive functions and brain functions therefore might require re-thinking the elemental functional units of cognition. New building blocks might well lead to a radically different overall picture of cognition, far away from the armchair. Instead of being restricted to dedicated phrenological circuits, the brain may be able to assemble different types of complex circuits from a variety of simple functional "lego"-blocks according to the behavioral demands.

The proposed research examines one potential building-block cognitive function across the boundaries of perception, working memory and long-term memory. The cognitive function is that of orienting attention voluntarily to a specific aspect of a representation. Attentional orienting is one of the basic executive control functions. It involves the ability to tune in at will to an anticipated aspect of stimulation. The best studied example is that of spatial orienting. If an event occurs at the expected location, we can detect it or make decisions about it much more quickly and accurately [4]. We can move attention to spatial locations independently from making eye movements.

So far, attentional orienting has been investigated within the perceptual domain, using largely a paradigm developed by Posner [4]. Essentially, an informative cue predicts the likely location of an upcoming event, which the participant must detect or identify. I hope to extend the investigation into the domains of working memory and long-term memory. The proposal introduces new experimental tasks in which informative retro-cues predict the likely location of an item in a past stimulus array or scene.

The rationale is simple. We optimize perception according to our expectations and what is relevant. It is likely that we can also optimize our access to mental representations in the same way, whether these are representations that we hold on-line from vanishing perception (working memory) or the that we summon from the slumbering memories formed in the past (long-term memory). Investigation of attentional orienting to these two types of mental representations as well as to perception will reveal whether neural circuits that achieve orienting are separate for the different domains or whether they share a common structure. The finding that this elementary function of cognitive control is not domain-specific will challenge the neo-phrenological dogma that cognitive domains have independent organization in the brain.

Mental representations and voluntary orienting

Investigation of attentional orienting to mental representations also provides a tool for probing the nature of mental representations per se. Mental representations are the templates of our internal mental world. Their nature is notoriously difficult to investigate, given the lack of direct or objective measures of their contents. Though mental representations may have their origins in the energies of the external world, they may also be highly constructed, abstracted or distorted interpretations. Furthermore, the nature of mental representations may change over time.

An event in the world is transformed into a perception. After it passes away, it is thought to leave a high-fidelity representation. In the case of a visual image, this is known as an iconic memory. In less than a second, the iconic memory either is lost or becomes actively maintained in consciousness as working memory. With time and interference, the working memory falls out of consciousness. At some later point, it may be possible to retrieve the event, whose representation has become consolidated in long-term memory

The details of how perceptions evolve into long-term memories are not at all well worked out. The favored idea is that there are different types of mental representations: iconic, working-memory, and long-term memory. For example, standard theoretical models propose separate neural stores for these different types of representations [6-7]. However, a simple and feasible alternative is that mental representations are formed and maintained within the brain areas directly involved in perception and action [7], without any need for additional independent stores. Representations and processing are bundled within the perceptual-motor information-processing systems much in the way as in connectionist networks [8]. Integration of distinct aspects of mental in working memory and in long-term memory appear to involve additionally associative brain areas, which do not store any information themselves, but which bridge between and coordinate activity in perceptual-motor systems [9-11 ]. In this view, mental representations may change gradually and continually over time through elaboration and consolidation mechanisms, perhaps without any major discontinuities.

Attentional orienting studies will first determine whether access to specific aspects of mental representations is under voluntary control in the same way as access to aspects of perceptual representations. Previous work has suggested that search for items in working memory proceeds through a rigid automatic serial and exhaustive process [12]. The selective control over access to long-term memory has not been tested directly. Finding that we can use informative retro-cues to enhance access to mental representations will be a breakthrough, showing that mental representations are flexible. Significant differences in the ability to orient attention to iconic, working or long-term memories would suggest different natures of these representations. In contrast, similar effects of attentional orienting would favor a common makeup. Brain imaging and recording experiments will reveal the functional anatomy and neural dynamics involved in accessing mental representations of different durations, to examine the degree of overlap versus specificity directly.

The success of the approach will open doors to investigating what types of predictive information facilitate access to memories of different durations (e.g., location, features, objects, object categories, etc), and therefore what types of information are coded and relevant in mental representations. Learning about the nature of mental representations will not only shed light on one of the longstanding mysteries about cognition, it may also have concrete practical implications for everyday life. We should be able to exploit voluntary orienting of attention to improve our mental capacity.

References:

1. Gross, C. G. (1992). Representation of visual stimuli in inferior temporal cortex. Philos Trans R Soc Lond B Biol Sci, 335(1273), 3-10.
2. Desimone, R. (1996). Neural mechanisms for visual memory and their role in attention. Proc Natl Acad Sci U S A, 93(24),13494-9.
3. Miyashita, Y. (1993). Inferior temporal cortex: where visual perception meets memory. Ann Rev Neurosci, 16, 245-63.
4. Posner, M. I. (1980). Orienting of attention. Q J Exp Psychol, 32(1), 3-25.
5. Waugh, N. C., & Norman, D. A. (1965). Primary memory. Psychological Review, 72(2), 89-104.
6. Atkinson, R. C., & Shiffrin, R. M. (1971). The control of short-term memory. Sci Am, 225(2), 82-90.
7. Crowder, R.G. (1993). Systems and principles in memory theory: Another critique of pure memory. In Theories of memory (Collins AF, Gathercole, SE et al., Eds). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc., pp. 139-161.
8. Rumelhart, D. E., & McClelland, J. L. (1986). Parallel Distributed Processing, I-II, Cambridge, MA: MIT Press.
9. Squire, L. R., & Kandel, E. R. (1999). Memory: from mind to molecules. Scientific American Library. New York: WH Freeman & Co.
10. Mesulam, M. M. (1994). Neurocognitive networks and selectively distributed processing. Rev Neurol (Paris), 150(8-9), 564-569.
11. Petrides, M. (1996). Specialized systems for the processing of mnemonic information within the primate frontal cortex. Philos Trans R Soc Lond B Biol Sci, 351(1346), 1455-1461.
12. Sternberg, S. (1966). High-speed scanning in human memory. Science, 153(736), 652-654.