Bookmark and Share

Alex Bäcker's Wiki / A neurobiological theory of autism
  • If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • You already know Dokkio is an AI-powered assistant to organize & manage your digital files & messages. Very soon, Dokkio will support Outlook as well as One Drive. Check it out today!

View
 

A neurobiological theory of autism

Page history last edited by Alex Backer, Ph.D. 11 years, 2 months ago

A Neurobiological Theory of Autism

Alex Bäcker

 

 

Autism is a developmental disorder, apparently caused by an injury at the time of closure of the neural tube (Rodier et al., 1996). Autistic subjects typically are usually asymptomatic until at least the age of two or so (Sacks, 2003). I suggest that this is because autistic subjects stay as infants, never able to learn to tell the important from the irrelevant (see also Snyder, 1997 and Hirstein et al., 2001). More precisely, I propose that autistics do not learn to categorize images and other sensory inputs using the invariances or equivalence classes that characterize an object class. Thus, their perceptual representations do not fall into the attractors that represent object classes.

 

This learning, deficient in autistic subjects, may be mediated by modification of feedback connections from so-called higher brain areas to lower sensory areas. Some of these feedback connections –those from the amygdala-- mediate the emotional significance of percepts. But not all of them do –autistic savants’ drawings are more realistic not because normal subjects pay attention to “emotional” parts of images, but simply because they categorize objects into a scene and then remember the categories, the concepts, rather than the actual stimuli. One reason to believe autistics’ learning deficit may be confined to corticofugal pathways is that, clearly, autistics are able to learn much, including the names of things (numbers, for example).

 

A deficit involving plasticity in autistics is supported by recent evidence that the most striking feature of autistics’ cerebrum is a much decreased rate of change compared to controls (Carper et al., 2002). More importantly, a recent study found a 30% reduction in cholinergic receptors in the cerebral cortex and basal forebrain (Perry et al., 2001). Nicotinic receptor frequencies are also modified in the cerebellum of autistic patients (Lee et al., 2002).The cholinergic system is involved in attention and memory (Everitt and Robbins, 1997).

 

This may explain why autistic savants’ skills do not improve with practice (Selfe 1977; O’Connor, 1989; Treffert, 1989; Sacks, 2003): because it is precisely the lack of learning the invariances and categories or clusters of natural objects that characterizes autistics.

 

What is central to the theory is not an impairment in plasticity per se, but rather an impairment in the ability for the gain of ascending synapses to be modified under top-down influence. These modifications may be mediated by synaptic modifications or by transient top-down activity, such as shunting inhibition.

 

Autistic savants are known to have outstanding skills of equipartioning time, space and quantity (Snyder and Mitchell, 1999). For example, some have the ability to keep time to the second (Treffert, 1989). Others can instantly tell the number of hundreds of matches fallen from a match box. Yet others can estimate the width of far-away objects with millimetrical precision. And others can calculate huge prime numbers with lightning speed, something that has been attributed to prime numbers standing out as “not capable of being equipartioned” (Snyder and Mitchell, 1999). We know that neuronal activity underlies behavior, and that individual neurons integrate across inputs from many other neurons. Normally, projections from different neurons have different synaptic weights, making the total activation of any neuron a poor indicator of the number of fibers activated, itself an indicator of the size of the external stimulus present. If autism impairs modification of the gain of individual ascending fibers, as I suggest, then synaptic weights, or more generally, gains, may well be more uniform than in a normal population, enabling a more direct relationship between neuronal activation and external stimulus quantity. In other words, I am suggesting that autistic savants’ abnormal ability to equipartion time, space and quantity is underlain by an equipartition in ascending fiber gain (or possibly synaptic strengths).

 

Simply put, autism may be what the brain does when it’s allowed to take the world in as is, so to speak, with no interpretations imposed on perception.

 

A peculiarity of autistic savant artists is that they often make left-right inversions/confusions (Sacks, 2003). At first sight, this seems to run contrary to their taking in images unmodified, uninterpreted. It seems like in that case, contrary to most, autistics do perceive an invariance, an invariance to lateral reflection. Yet Wetzel and Bäcker previously showed that left-right invariance is innate, not constructed, and is in fact lost as we grow up, further suggesting autism as a developmental disorder. Furthermore, there is a way in which this invariance is different from the rest: left-right comparisons depend on comparisons across hemispheres[1]. On corticofugal projections across hemispheres –since the commisure linking hemispheres at the level of the thalami, the MASSA INTERMEDIA, is absent in many normal brains[2]. It may thus not be surprising that left-right comparisons are impaired in autistics. But what is really going on when autistic savants flip an image remains a mystery.

One prediction of this theory is that autistics should be deficient in tasks requiring top-down processing. It is striking how few histological studies have been performed on the brains of autistic patients[3]. An interesting recent study, for example, showed that nicotinic receptor frequencies were modified in the cerebellum of autistic patients (Lee et al., 2002), but I am aware of no equivalent studies in the cerebrum.

 

An interesting corollary of this theory, if correct, is that autistics would prove a model system to figure out the long elusive functional role of corticofugal projections, which are known to be massive (see Koch, 1987). The considerations above suggest a role of corticofugal projections modifying the gain of individual ascending fibers according to the corresponding percept’s “compatibility” with the attractor closest to the aggregate activity, i.e. the representation of the activated concept or category. Corticofugal pathway would provide the context and interpretation and modify sensory inputs even as they enter the cortex, explaining the pervasiveness of neuronal modulations due to illusions throughout visual cortex. This proposal is consistent with a previous proposal that corticofugal input serves to control the gain of retinogeniculate synapses (Koch, 1987).

 

An alternative explanation for autistic savants and genius in general, due to Ramachandran (1999), suggests that they have overdeveloped and hypertrophied a single part of the brain at the expense of others, and that the extra neuronal hardware gives them skills normal subjects do not have. There may  be truth to the contention that increased neuronal mass allows an increase in skills underlain by the area in question. But given the huge variability in brain mass among geniuses, it seems hard to believe that small absolute increases in a brain area’s mass underlie such phenomenal skills. More importantly, Ramachandran’s hypothesis predicts that the focussed and compulsive practice of a single skill that autistics usually exhibit would further enhance the relative brain mass devoted to that task, and thus further enhance the corresponding skill. And yet the evidence suggests instead that autistic savants’ skills arise suddenly and do not improve with practice (Selfe 1977; O’Connor, 1989; Treffert, 1989; Sacks, 2003), as has been pointed out also by Snyder and Mitchell (1999).

 

Spatially localized inactivation experiments could help elucidate whether autistics’ deficit is specific to corticofugal projections or extends also to gray matter in higher cortical areas, and whether loss of growth in gray matter is a primary event or one secondary to attrofiation of their projections.

 

Alex Bäcker, Ph.D. 

Pasadena, 2003

 


[1] Incidentally, this may also be why children normally take so long to learn left and right: they may need to learn to integrate across both of their hemispheres –which we know, thanks to the work of Sperry, Bogen, Gazzaniga and others, can function quite independently-- to do this.

[2] http://www.driesen.com/glossary_m-n.htm#MASSA%20INTERMEDIA .

[3] A Web of Science search for autis* and histol* yields a mere 16 articles, none older than 1996.

Comments (0)

You don't have permission to comment on this page.