Monday, 8 April 2019

A speculation: moving from one tooth to the next

Figure 1
We suppose that we are gazing at a long toothed bar, roughly as above. The task is to move one’s attention in an orderly way from left to right and readers are invited to try this for themselves, possibly clicking on the image to enlarge it first. We find that we can do the whole bar in around 20 seconds, a rate of something better than 1Hz and a process helped along by nodding slightly and counting in time. The brain seems to like, to respond to simple rhythms which are expressed physically.

As far as consciousness is concerned this process might be expressed in one of the following ways.

Option 1: lock on to a tooth and hold that lock for a couple of hundred milliseconds or so; release attention and move eyes to the right; twist head to the right; lock on to the next tooth; release attention and so on. The twisting head bit probably being at the border of consciousness: one is aware after a while that the head has moved, but one is probably not conscious of each small, catching-up-with-eyes increment.

The problem is that given that the brain can only consciously attend to one thing at a time, how does it jump to the next thing? How does it know, how is it sure, for example, that the thing that it has jumped to is the next thing?

Option 2: make a tooth the object of attention, expand the object of attention to the right to include the next tooth, contract the object of attention from the left to exclude the first tooth, make the next tooth the object of attention. From where we associate to the way that a worm moves along the ground, with successive contractions and expansions of its body.

This formulation pushes the business of moving eyes and head about into the background and concentrates more on the subjective experience.

Figure 2
The idea is that all the teeth are much the same and that there are no good clues about position on the bar from the context. For example, the red dots on alternate teeth, as in the figure above, are a great help to the task. Instead, the attended to tooth is the one on the foveae (roughly at the centre of the retinas) when we look straight ahead, when the eye balls are properly lined up with the nose, the ears and the shoulders, seemingly the natural, the usual position when we attend to something, as can be seen, for example, in pointing dogs. We suppose that the brain knows when the eyes are looking straight ahead, perhaps helped along by the power of stereoscopic vision.

We have talked in the past of various counting tasks – see, for example, references 5 and 6 – and we will talk, in a post to come, more generally of some of the many different kinds of counting tasks, mostly visual, and the various difficulties encountered therein. Here we concentrate on this one.

Figure 3
We have added a virtual red spot to the snap above, to indicate the tooth that is presently being attended to. Then, although we are attending to the red spotted tooth, the eyes are moving around the whole time, lots of saccades, in the jargon of eye tracking people – for whom see, for example, reference 1. Saccades which, during this task, tend to be small but which can be large. With part of the purpose of these saccades being to maintain the stimulation of the retina, to maintain the rate of change on the retina, without which the conscious image will fade. On which point see, for example, reference 2.

We then suppose that the brain is randomly generating saccades along the bar, giving us something like a one dimensional random walk. But there are biases. First, the eyes tend to jump to the right so as to reconnoitre the way ahead. Second, the eyes keeping jumping back to the tooth that is being attended to, or teeth in the case of the second option above, the one in each of the two foveae when looking straight ahead. The eyes need to spend quality time on the tooth which is the subject of conscious attention in order to keep the subjective experience up to strength, up to snuff.

After each saccade, the brain decides whether it is looking at tooth or not tooth.

Figure 4
After a while, that is to say after a few hundred milliseconds, it has built a map of the bit of bar around the tooth currently being attended to, perhaps in the form of a histogram like that suggested above, with the height of the blue bars suggesting the number of hits and the pink bars the number of misses. A histogram which will contain noise but which contains enough information for present purposes, and it is reasonably straightforward to identify the next tooth to the right of the current tooth and to compute the saccade needed to get to it.

After a further while, the brain might decide that the teeth are arranged in an orderly way and need less additional input to compute the next saccade.

All this is happening more or less subconsciously, with the conscious attention remaining locked on to the tooth which we have marked with the red spot.

Implications for LWS-N

The LWS-N proposition – for which see reference 3 - is that consciousness is organised into a succession of frames, each lasting of the order of a second or so. Frames can be updated, but for big updates the brain starts over. We talk of frames being compiled, by analogy with the computer programs of the 1980’s.

The visual part of consciousness will be held on one or more layers of LWS-N, with the teeth of the present note probably be represented as a mixture of layer objects (the current teeth, the important teeth) and parts of the layer objects (the rest).

We had thought that the object being attended to will be marked, inter alia, by a column object linking the visual part of that attention to other parts.

As our attention moves from one tooth to the next, at the very least that column object is going to have to be moved and during that movement, while there may be attention, there will not be a single tooth which is the object of that attention.

Figure 5
However, the current thinking is that the object being attended to will be marked by the activation process, with the unit of activation being either a layer object or one part of a layer object. Attention had perhaps moved from being a binary thing, present or absent, to something more graduated.

In this case, provided there is enough information in the visual layer, the shift from tooth to tooth can be managed by shifting the activation processes, as suggested in the figure above. In the middle of this process, the visual detail of the teeth may be weakened, with the subjective experience focussed just on the fact of there being two teeth and the need to move from one to the other. From where we associate to Tononi and Koch’s talk at reference 7 of our experience being both integrated and differentiated, which may, in part, be grasping at the same problem.

All this being implemented by large numbers of neurons firing away, possibly as many as hundreds of thousands of them, under the hood, as it were.

From time to time, the visual layer will need to be refreshed, a new frame will need to be built and it is, perhaps, at these times that it is going to be particularly easy to lose one’s place in the row of teeth.

Testing

Figure 6
Our understanding is that it would not be difficult to test the eyes part of all this, in that using the eye trackers already mentioned and some sort of experimental set up, with a computer screen to look at and with some buttons to press, one could observe what the eyes were doing and relate that to what subjects are experiencing. While Dehaene and his colleagues would no doubt hook into an EEG machine, also to be bought from the people at reference 1. For Dehaene, see reference 8.

With the figure above suggesting the sort of thing that eye trackers can produce, taken, via Wikipedia, from reference 4. ‘This study by Yarbus (1967) is often referred to as evidence on how the task given to a person influences his or her eye movement’. No doubt things have moved on since then.

Testing of LWS-N being rather more difficult, our not yet having any very clear idea of where it is – quite apart from the difficulty of recording small electrical affairs deep inside the living, human brain.

Conclusions

We have offered a speculation about how the brain, while only attending to one thing at a time, can jump in an orderly way from one thing to the next.

References

Reference 1: https://imotions.com/.

Reference 2: The fading of stabilized images: Eye movements and information processing - Stanley Coren, Clare Porac – 1974.

Reference 3: http://psmv3.blogspot.com/2018/05/an-update-on-seeing-red-rectangles.html. A slightly out of date portal to the world of LWS-N.

Reference 4: Eye movements and vision - Yarbus, A. L. – 1967.

Reference 5: https://psmv4.blogspot.com/2019/03/counting.html.

Reference 6: https://psmv3.blogspot.com/2018/09/shopping-lists.html. Second part is about counting the floors of Vauxhall Tower.

Reference 7: Consciousness: here, there and everywhere - Tononi & Koch – 2015.

Reference 8: Consciousness and the Brain: Deciphering How the Brain Codes Our Thoughts - Stanislas Dehaene – 2016.

Group search key: srd.

No comments:

Post a Comment