8/22/2014

Eye movements in ADHD: Not a "foolproof" diagnostic method, but interesting and important

Above: Eye movements. Are they really a foolproof clue to ADHD?

Science Daily claims that measuring "Involuntary movement [is] a foolproof indication for ADHD diagnosis." Specifically, they believe it will reduce the rate of misdiagnosis in children:

Attention deficit hyperactivity disorder (ADHD) is the most commonly diagnosed -- and misdiagnosed -- behavioral disorder in American children. Now a new study can provide the objective tool medical professionals need to accurately diagnose ADHD. The study indicates that involuntary eye movements accurately reflect the presence of ADHD.

Unfortunately, the actual study, by Fried and colleagues1, indicates nothing of the sort.2

The study did, in fact, measure involuntary eye movements--specifically, blinking and "microsaccades," small jerky eye movements.

22 adults with ADHD took the TOVA twice, the first time unmedicated and the second time while taking methylphenidate (Ritalin).2  The TOVA, which stands for Test of Variables of Attention, is a common, computerized diagnostic test for ADHD.  While participants took the test, their involuntary eye movements were measured.  The unmedicated ADHD adults made more microsaccades and blinks than neurotypical adults immediately before the onset of the stimulus.  When they took the test a second time on methylphenidate, they did not differ from the control group. The researchers argued that measuring involuntary eye movements is more precise, quantitative, and harder to "game" than many current diagnostic methods, such as questionnaires and self-report.

Unfortunately, this study cannot possibly generalize enough to a wider population to imply anything about diagnosing ADHD in the real world.  First of all, although the Science Daily press release focused on children, and parents' concern about their taking stimulants unnecessarily, Fried's study only examined adults.  Furthermore, the study measured eye movements only during a single specific task, the TOVA.  It's not clear whether the same finding would occur during other tasks, or even at rest.  And of course, a sample of 22 participants may not generalize well to the millions of people with ADHD.

And even if the findings generalize to other ages and tasks, further conditions would need to be met to ensure the method worked in practice.  It would have to be tested in a real-world school or clinical setting, with a large population of people referred for possible ADHD (on the order of hundreds or even thousands).  It would have to successfully distinguish ADHD not only from typical development but from other forms of atypical development--which is much harder.  While most of the hysteria about ADHD misdiagnosis assumes that the misdiagnosed children are "normal" children treated as if they have a developmental disorder, it's probably more common for children to be misdiagnosed as having ADHD because of real symptoms caused by mental illness (such as depression), physical illness (such as sleep problems) or another developmental disability. Furthermore, the method would have to be shown to be cheap, fast, and practical in real-world contexts.  For example, the TOVA is a common diagnostic test among neuropsychologists and other specialists, but to the best of my knowledge, it is less so among pediatricians who may also be called on to diagnose ADHD. If the TOVA is essential to obtaining this pattern of eye movements, the method may not generalize well to all diagnostic settings.

This is not even the first study to identify increased anticipatory saccades as a marker of ADHD.  In a review, Rommelse and colleagues report five others (Castellanos et al., 2000; Mostofsky et al. 2001a & 2001b; Rommelse et al., 2008, & Ross et al., 1994; see reference list).  Bittencourt and colleagues' review described a study by Feifel and colleagues, who tested ADHD adults who went unmedicated at least 48 hours. These participants generated more anticipatory saccades when a task required them to shift their attention from central fixation to a target appearing randomly onscreen.

Furthermore, Richard N. Blazey, David L. Patton, and Peter A. Parks have a U.S. patent for a method of detecting ADHD through saccades, or what they call "angular movements of the eyeball."  Their patent differs from Fried's method in one crucial respect.  Eye movements are measured while the person sits staring at a blank screen, wearing noise-canceling headphones, rather than while doing a demanding test like the TOVA.

So Fried's study doesn't offer a foolproof way to diagnose ADHD, and measuring unconscious eye movements isn't a new way to diagnose ADHD in the first place.  But the paper still offers an important insight, one left unexplored by the Science Daily article.  ADHD adults make more anticipatory saccades than neurotypical adults, and it turns out that anticipatory saccades play an important role in perception and tell us something about the strengths and weaknesses of vision in ADHD.

In order to understand why the finding matters, let's briefly go over the research on what anticipatory saccades do and why they matter.

First of all, a saccade is a rapid, jerky eye movement that brings something new into focus on the fovea, the most sensitive part of the eye.  It is not the only sort of eye movement (there are also smooth pursuit eye movements that smoothly track a moving object already in focus).  However, it is extremely common and essential to vision.  As Michael F. Land explains:

“Throughout the animal kingdom, in animals with as diverse evolutionary backgrounds as men, fish, crabs, flies, and cuttlefish, one finds a consistent pattern of eye movements which can be referred to as a ‘saccade and fixate’ strategy. Saccades are the fast movements that redirect the eye to a new part of the surroundings and fixations are the intervals between saccades in which gaze is held almost stationary. As Dodge showed in 1900, it is during fixations that information is taken in: during saccades we are effectively blind.” 
Saccades can take in varying amounts of the visual field (measured in degrees).  Microsaccades, like the ones in the Fried ADHD study, are generally defined as extending to less than 15 minutes of arc in the visual field.

Even a cursory Google search reveals that anticipatory eye movements play an important role in a variety of tasks at all ages.  Anticipatory eye movements have been used to study expertise in soccer goalkeepers, object perception and expectations for "dynamic visual events" by babies, discourse comprehension in typical adults, sentence processing in adult readers, and the influences of language and knowledge about the world on adults' real-time visual attention.  These lines of research reflect an understanding that anticipatory eye movements are overt indicators of otherwise hidden attention shifts.  It's worth noting that not all anticipatory eye movements are microsaccades like the ones in Fried's ADHD study.  Some are smooth pursuit eye movements (the smooth, rapid eye movements used to track moving objects over a longer period of time).  However, saccades are often used for anticipatory eye movements, and the two overlap in neural areas, task demands, and the cognitive processes believed to be involved.

As Ellen Kowler put it:


"Work over the last 25 years has also converged on the notion that the saccadic system is inherently predictive, using pre-saccadic shifts of attention and signals representing planned saccades to encode the location we are about to fixate, and to prepare visual neurons, in advance, for the post-saccadic image. These predictions may be instrumental in processes ranging from the control of saccadic accuracy (by means of adaptive saccadic adjustments) to the weaving together of discrete glances in a way that gives us the impression of a clear and stable perceptual world despite the continual displacements of the retinal image produced by saccades."
Anticipatory or "look ahead" saccades occur during a variety of real world tasks, including text and sheet music reading, typing, looking at pictures of scenes, drawing, walking over difficult terrain, driving, sports, and cooking. In reading, saccades select a span of 7-9 letters or 1-3 notes for processing.  Typists keep their eyes about 1 second ahead of the currently typed letter.  Sketching involves rapid cycles of gaze shifts between the person being drawn and the drawing.  People instructed to walk in specific, effortful ways will look ahead by 1 to 2 steps.  When steering around a turn, driving instructors look at points their car will not reach for another 2-3 seconds, whereas their students, who have not yet learned to anticipate this way, look straight ahead instead.  Tennis players' saccades anticipated the bounce of a tennis ball by about 0.2 seconds, while cricket players can anticipate the ball by about 0.1 seconds.  When making tea or a sandwich, at the end of each action, the eyes move on to the next object about half a second before beginning the next action.  When there are two objects involved (e.g., a kettle and lid), multiple saccades and fixations must be made between the two objects.  In general, saccades followed by fixations proceed a step ahead of action.  This avoids reliance on visual memory.

All this anticipation is believed to help people do things better.  For example, they help people aim their reaching and grasping movements accurately.  Crucially, these saccades may improve performance on visual tasks and arm movements even when attention is directed elsewhere.

In general, we see, hear, act on, and think about things better when we are attending them.  Therefore, the visual system needs to predict what objects or locations in the visual field will be important and start attending to them before they do anything important.  Anticipatory eye movements are the mechanism for doing this.  As such, extra anticipatory eye movements should be an advantage.  They should increase the likelihood of attending the right thing and perceiving it accurately.  Or, at least, that is the picture presented by research on vision in the general population.

ADHD research describes anticipatory eye movements quite differently.  It describes them as "premature," "impulsive," and a sign of inhibitory difficulties--even when the tasks being studied do not involve inhibition of any kind.

So, is there any real difference between "premature saccades" and "anticipatory" ones, or are they just a negative and a positive label for the same behavior? I asked Sue Fletcher-Watson about this, and it appears that there is a difference, which is only apparent in studies designed specifically to test learning visual patterns. In studies like this, anticipatory saccades reflect an accurate prediction of where a stimulus will appear, based on learning; premature saccades are the eyes' tendency to flick around randomly looking for stimuli when a person gets bored.  Unless a person making premature saccades is lucky, their saccades will have no tendency to land where a new stimulus will appear, so the "success rate" should be measurably lower for premature saccades than truly anticipatory ones.

However, many of the ADHD studies under discussion do not appear to use this sort of learning task. Instead, they measure individuals' control over their eye movements during an experiment that is static and does not encourage or require learning. I suppose individuals could learn in such studies, and their learning could be measured and analyzed, but that's not what these researchers were doing. They might need to reanalyze their data to measure and compare learning in typically developing and ADHD groups. (And perhaps someone should).

Furthermore, I don't think that the ADHD researchers were even asking that question, trying to determine whether ADHD behavior was anticipatory or just premature.  Instead, they seem to be simply assuming "ADHD is a disorder of inhibition, so group differences must be due to an inhibition deficit."  This is an example of biased research, as explained by Morton Ann Gernsbacher--if you take away the labels of the group members, the interpretation no longer makes sense.  It also fails to take into account the more positive picture of anticipatory eye movements painted by the general vision literature.

Researchers on atypical development often say we should take typical development as a starting point, so let's do just that and see what it actually implies for vision in ADHD.

Let's say for the sake of argument we reanalyze the data or do new studies to make sure the eye movements actually are anticipatory--and they are.

Increased anticipatory eye movements in ADHD could have positive effects, which could be viewed either as a simple advantage or as compensation.  The trait may also carry a disadvantage.

When viewed as an advantage, it implies that people with ADHD may be better at noticing and directing their attention to changing and salient things in their environment.3  This ties in well with the theory (described to me by Jeff Gilger's research team) that ADHD vision is well-adapted to dealing with rapidly-changing, highly-salient stimuli, but ill-suited to maintaining attention on static and less motivating stimuli.

When viewed as compensation, extra anticipatory saccades could be seen as a way of increasing the likelihood of focusing on the right things when attention is hard to control by other means.  Anticipatory eye movements could also compensate for lack of visual working memory.  In the general population, people who have to copy drawings or models make lots of saccades to the thing being copied, instead of looking back at the object less and relying mostly on visual memory.  Since children and adults with ADHD may have less visual working memory than neurotypical people, they may need to make even more saccades.

However, there may be a disadvantage to the ADHD pattern.  Remember that we are effectively blind while making saccades or blinking.  The more saccades and blinks a person makes, therefore, the less time they spend taking in visual information from their environment.  This can be a disadvantage when performing a task that requires taking in lots of visual information over a long period of time (e.g., certain sustained attention tasks).

These interpretations of Fried's study are merely my speculations, based on putting aside assumptions about ADHD and instead looking at what we know about how vision works, in general.  So please don't quote them as if they were established research.  Rather, these represent a possible blind spot in ADHD research--and a set of hypotheses we can test.  Even if they turn out to be wrong, we will learn something about the strengths and weaknesses that come with ADHD.

Many thanks to Sue Fletcher-Watson for helping me figure out how to tell the difference between anticipatory and premature saccades and offering encouragement.

1 Interestingly, Moshe Fried himself has ADHD.
2  At least, so far as I can tell from the abstract and media descriptions from the study. This post will be updated when I get access to the full text.
2 Notice the obvious confound: improved performance on the second testing could have been due to practice taking the test rather than the stimulant medication.
3 Note that having an advantageous pattern of anticipatory saccades, which are unconscious and not deliberately controlled, does not preclude difficulty with more controlled eye movements.  In fact, a line of research indicates that ADHD people have difficulty controlling and inhibiting eye movements, just as they do with other behavior.  ADHD could simply involve deficits in controlled, but not necessarily uncontrolled, perception and behavior4.
4 Whether or not uncontrolled ("reflexive") saccades are also impaired in ADHD is still under debate.  Some studies seem to find slower, more variable responses, while others observe no difference from neurotypical peers.

Citations:
Note:  The basic point about the strengths and weaknesses of anticipatory eye movements in normal development can be found in James Enns' The Thinking Eye, The Seeing Brain (one of the few readable, layperson-friendly textbooks I have ever seen).

Other citations reflect a limited amount of time to survey the role of anticipatory eye movements in ADHD and typical development, so they are mostly reviews and are necessarily incomplete.

  • American Friends of Tel Aviv University (2014). Involuntary eye movement a foolproof indication for ADHD diagnosis. Science Daily, 13 August 2014. www.sciencedaily.com/releases/2014/08/140813131055.htm 
  • Moshe Fried, Eteri Tsitsiashvili, Yoram S. Bonneh, Anna Sterkin, Tamara Wygnanski-Jaffe, Tamir Epstein, & Uri Polat (2014). ADHD subjects fail to suppress eye blinks and microsaccades when anticipating visual stimuli but recover with medication. Vision Research 101, pp. 62-72. http://www.sciencedirect.com/science/article/pii/S0042698914001187
  • Juliana Bittencourt, Bruna Velasques, Silmar Teixeira, Luis F. Basile, Jose Inacio Sailes, Antonio Egidio Nardi, Henning Budde, Mauricio Cagy, Roberto Piedade, & Pedro Ribeiro (2013). Saccadic eye movement applications for psychiatric disorders.  Neuropsychiatric Disease and Treatment, 9, pp. 1393-1409. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3783508/
  • F.X. Castellanos, F. F. Marvasti, J. L. Ducharme, J. M. Walter, M. E. Israel, A. Krain, et al. (2000). Executive function oculomotor tasks in girls with ADHD. Journal of the American Academy of Child & Adolescent Psychiatry, 39, pp. 644-650.
  • D. Fiefel, R.H. Farber, B. A. Clementz, W. Perry, & L. Anilo-Vento (2004). Inhibitory deficits in ocular motor behavior in adults with attention-deficit/hyperactivity disorder.  Biological Psychiatry, 56:5, pp. 333-339. http://www.ncbi.nlm.nih.gov/pubmed/15336515/
  • Eileen Kowler (2011). Eye movements: The past 25 years. Vision Research 51:13, pp. 1457-1483 http://www.sciencedirect.com/science/article/pii/S0042698910005924 
  • Michael F. Land (2006). Eye movements and the control of actions in everyday life. Progress in Retinal & Eye Research, 25, pp. 296-324. http://www.cis.rit.edu/pelz/scanpaths/papers/eye-movements-every-day-life-land-2006.pdf
  • Neil Mennie, Mary Hayhoe, & Brian Sullivan (2006). Look-ahead fixations: Anticipatory eye movements in natural tasks. Experimental Brain Research http://www.ski.org/Renninger_Lab/BSullivan/MennieEtAl_LookAheadFixations2006.pdf
  • S. H. Mostofsky, A. G. Lasker, L. E. Cutting, M. B. Denckla, & D. S. Zee (2001a). Oculomotor abnormalities in attention deficit hyperactivity disorder: A preliminary study. Neurology, 57, pp. 423-430.
  • S. H. Mostofsky, A. G. Lasker, H. S. Singer, M. B. Denckla, & D. S. Zee (2001b). Oculomotor abnormalities in boys with Tourette syndrome with and without ADHD. Journal of the American Academy of Child & Adolescent Psychiatry, 40, pp. 1464-1472.
  • Nadia N. J. Rommelse, Stefan Van der Stigchel, & Joseph A. Sergeant (2008). A review on eye movement studies in childhood and adolescent psychiatry. Brain & Cognition, 68, pp. 391-414 http://www.fss.uu.nl/psn/web/people/personal/stigchel/rommelsereview.pdf 
  • N. N. J. Rommelse, S. Van der Stigchel, J. Witlox, C. J. A. Geldof, J. B. Deijen, J. Theeuwes, et al (2008). Deficits in visual spatial working memory, inhibition, and oculomotor control in boys with ADHD and their non-affected brothers. Journal of Neural Transmission, 115, pp. 249-260.
  • R.G. Ross, D. Hommer, D. Breiger, C. Varley, & A. Radant (1994). Eye movement task related to frontal lobe functioning in children with attention deficit disorder. Journal of the American Academy of Child & Adolescent Psychiatry, 33, 869-874.
  • Alexander C. Schutz, Doris I. Braun, & Karl R. Gegenfurtner (2011). Eye movements and perception: A selective review. Journal of Vision 11:5, http://171.67.113.220/content/11/5/9.full.