As you probably know, saccades are to thought to be preded by attention shifts (see especially the work form Heiner Deubel). So you could assume in most cases, that saccade landing points indicate some form of attention shift. But as you already said, that is not always the case - but probably the best predictor if you want to observe free-viewing behavior.

In a fixation task, pupil size can be used (see e.g. Mathot et al. The Pupillary Light Response Reveals the Focus of Covert Visual Attention. PLoS ONE, 2013, 8, 1-10)

Or you could use microsaccades (see e.g. Hafed & Clark. Microsaccades as an overt measure of covert attention shifts. Vision Research, 2002, 42, 2533-2545).

Hope that helps...

Bear in mind that covert shifts of attention can happen in the absence of eye movements (which is why they're called covert).  See for example the introduction of the article linked below.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2376252/

Thanks for the responses David and Bruce, both very helpful. I will take a look at the references you've highlighted.

Hi Jim,

This is a very interesting topic. Covert attention has been studied a lot in laboratory settings, when eye movements are explicitly controlled but we know very little on the role covert shifts of attention play in real-life situations. Some have argued that they may play no role (e.g. work of John M Findlay). I think it might be true for spatial shifts of attention, but other attentional phenomena are certainly important (so-called feature-based or object-based attention), as the inattentional blindness example illustrates. 

I am not sure that you will be able to use eye movements themselves to determine what the subjects actively see. Of course it depends on what you are trying to do exactly but I would suggest rather using some sort of concurrent perceptual task to probe attention. It's very tricky but when done carefully it can work very well. 

Alex

There is evidence that during reading it's possible to use eye-tracking to determine episodes of mindless reading, i.e., periods when readers don't pay attention to the text:

Schad, D. J., Nuthmann, A., & Engbert, R. (2012). Your mind wanders weakly, your mind wanders deeply: objective measures reveal mindless reading at different levels. Cognition, 125(2), 179-194.

Here, the observation was the following: During reading, fixation durations reflect ongoing word processing: words that are difficult to read (e.g., low frequency) are fixated longer than easy words (high frequency). When the mind wanders away from the reading task during mindless reading, text processing is reduced and fixation durations are no longer coupled to text difficulty. Based on this effect, it is possible to detect episodes of mindless reading from eye-movement by identifying time points when fixation durations don't match the fixated words' difficulty: e.g., when readers fixate a very difficult low frequency word, but fixate it only for a very short time, this can provide an index for mindless reading (i.e., not paying attention to the text).

To my knowledge, in how far this logic can be applied to other tasks needs to be tested..

Article Your mind wanders weakly, your mind wanders deeply: Objectiv...

Thanks for the responses Alexandre and Daniel. I've not been able to find much in the literature about eye movements in 'real', everyday situations and what they reveal about attention. The research around eye behaviour during reading and mind-wandering is really interesting (thanks for the ref Daniel) and is what makes me wonder if similar patterns of eye movements /  fixations can be used to determine mind-wandering or locus of attention in other settings. In a previous study I have tried to use an ongoing secondary perceptual task alongside eye-tracking to identify moments when attention may have been directed to what was being looked at (when performance on the secondary task reduced below a mean threshold - link to paper is below). This seemed to work to a degree but had a number of limitations so I'd be interested in finding other methods of identifying when and where attention is directed.

http://lrt.sagepub.com/content/early/2014/05/09/1477153514522472.1.full.pdf+html

Hi Jim,

may be I could help with literature about real world tasks and eye tracking:

Hayhoe, Mary M.; Rothkopf, Constantin A. (2011): Vision in the natural world. In: WIREs Cogn Sci 2 (2), S. 158–166. DOI: 10.1002/wcs.113.

Land, Michael F.; Tatler, Benjamin W. (2009): Looking and acting. Vision and eye movements in natural behaviour. Oxford, New York: Oxford University Press.

Rothkopf, Constantin A.; Ballard, Dana H.; Hayhoe, Mary M. (2007): Task and context determine where you look. In: Journal of Vision 7 (14), S. 16. DOI: 10.1167/7.14.16.

Tatler, Benjamin W.; Hayhoe, Mary M.; Land, Michael F.; Ballard, Dana H. (2011): Eye guidance in natural vision: reinterpreting salience. In: Journal of Vision 11 (5), S. 5. DOI: 10.1167/11.5.5.

Torralba, Antonio; Oliva, Aude; Castelhano, Monica S.; Henderson, John M. (2006): Contextual guidance of eye movements and attention in real-world scenes: The role of global features in object search. In: Psychological Review 113 (4), S. 766–786. DOI: 10.1037/0033-295X.113.4.766.

Best regards,

Matthias

Yes, you can, but only on average over many trials.

Specifically, from the orientation (not direction!) of microsaccades, you can infer the radial position of the focus of spatial/object attention.

The catch is that microsaccades are rare (particularly when the information displayed is task-relevant) and don't happen on every trial.

To be able to average over trials, you need to ensure that attentional deployment is reproducible.

Best regards,

Jochen

Refs: Pastukhov, Vonau, Stonkute, Braun (2012), Vision Research, Spatial and temporal attention revealed by microsaccades.  http://dx.doi.org/10.1016/j.visres.2012.11.004

Yes, I am perfectly sure that eye movements are correlated with the processing of tactile or visual stimuli. More interestingly, it was even reported that microsaccades reflect the categorization of auditory stimuli. 

http://www.ncbi.nlm.nih.gov/pubmed/25122911

cheers

Tobi

Also another point I forgot to make is that we found recently that when you have stimuli in motion, eye velocity and acceleration seem to match the velocity and acceleration of the stimuli on which attention is allocated, even when the eye position does not match the stimulus location (i.e. covert attention is used). I guess it could be useful to your application if you have dynamic visual stimulation. It's unpublished data though. 

Thanks all for the insightful and interesting answers you've provided, it's given me plenty of food for thought, and plenty of references to pursue.

Eye movement alone tells you very little. When engaged in concentrated thinking, people move their eyes away from any distracting object, e.g., faces, and eyes remain stationary.  Similarly, processing of visual stimuli is associated with a fixed gaze - processing of visual input is best indexed by pupilometry -- on which there is a large literature.  

With respect to Bruce's comment. There is a covert processing system! We can attend to what's going on in peripheral vision without looking at it for sure but also what occurs in peripheral vision as actually temporally in advance of or temporally prior-to what goes on in central vision. The holistic form of attention in peripheral vision 'posts' what is considered important to central vision so that it can attend to it. Decisions have even been made subconsciously. Not only that the lens is changing its shape to get ready for the required focal length of the subject of next point of attention before saccadic eye movement begins. So peripheral vision is spatially salient / cognisant (even on a monocular basis)? Too much to go into here but attached a list of presentations that you may find of interest.