What is attention?

Good science is supposed to be "frugal," i.e., it ought to make use of as few assumptions and abstractions as possible. In neuroscience, the abstract concept of "attention" is a concept that is considerably less frugal than would be desirable. It is basically a black box and does not necessarily explain which processes in the brain it actually addresses -- a central question of perception research today.

For several decades, "attention" was thought to be a barely definable state of certain brain regions. In visual perception, for instance, eye movements towards a stimulus are triggered in the Superior Colliculus, a part of the midbrain. The direction of attention in the brain does not react equally to all stimuli, though: when there is a high level of "attention in the sensory system, reactions are swift and intense; neuroscientists call this state "attentional capture." A state of slow and comparatively weak reactions, on the other hand, is called "inhibition of return" (IOR). Attentional capture and IOR both follow an alternating pattern, which rides on a rhythm with approximately 10 oscillations per second.

But what causes this rhythm, and how does "attention" control it? The international research team have now been able to depict the processes which might be responsible in a surprisingly simple model. These CIN and NIPS neuroscientists have been cooperating for several years now, headed by Dr. Ziad Hafed (Tübingen). They are investigating a phenomenon whose importance for visual perception has long been underestimated: tiny eye movements, so-called microsaccades. These small adjustments constantly correct the axis of vision when the gaze fixates an object. In earlier studies, Hafed and colleagues had found that microsaccades are generated in the Superior Colliculus in a stable rhythm that is reset by new stimuli entering the visual field. Microsaccades also change direction with each iteration.

Following up on these observations, Hafed and his team arrived at a hypothesis: what if the rhythm and direction of microsaccades directly trigger attentional capture and IOR? They developed a theoretical and computer model based on this assumption, employing a wide range of parameters to see what predictions the model could make. Testing the model's predictions in experiments, the researchers surprised even themselves: they found that besides microsaccades, no additional factors were necessary in the model to explain attentional capture and IOR.

Ziad Hafed states that "attention" may be explained quite "parsimoniously." He believes that the brain filters "important" stimuli simply based on saccadic corrections of the direction of gaze. These eye movements directly produce the phenomena which have so far been described as attentional capture and IOR. Which of these two occurs depends on the timing and direction of the stimulus relative to the microsaccadic rhythm and direction. 'These findings are a strong argument that the mechanism of "attention" is based on a very simple principle', says Hafed. 'Should they be validated by further studies, results from decades of research would have to be seen in a very different light.'