Developmental paediatrics: January 2020

Sunday, 26 January 2020

Embodied cognition

I tried turning it off and then on again already!

'Embodied cognition' is cognition which is "deeply dependent upon features of the physical body of an agent, that is, when aspects of the agent's body beyond the brain play a significant causal or physically constitutive role in cognitive processing".

It provides an alternative way of understanding the human mind, to the prevailing cognitive science model. In the cognitive science model, the brain is like a computer, it receives information from its external senses, it processes that information, then it provides a (motor) output. It processes information through computation. Its job could be entirely performed instead by a (very powerful) computer).

I myself find a lot of embodied cognition theory and research compelling and illuminating, it feels closer to real-life to me than some cognitive science. At present, embodied cognition is a minority view, and it's also not a single group or research program, it covers many different groups, with many different (and sometimes conflicting) theories.

It can be tricky to describe, so let's take a look at a couple of examples.

*****

The outfielder problem

I'm not saying I entirely understand this diagram...

In cricket*, how does an outfielder know where to run, to catch a ball which has been hit high and far by the batsman/batswoman?

'Instinct'. 'Practice'. Yeah sort of, but I mean the deeper explanation. How does the brain work out exactly where to send its body, in time to catch the ball?

Given the ball's starting point, and an estimate of its direction and velocity soon after it was hit, the answer can be calculated, it's basic Newtonian mechanics. Does the brain do this? Does it turn the two-dimensional image which the eyes have sent it into a 3D model, calculate the ball's landing spot, and send the body there? This model has been proposed as an explanation.

But that's apparently not what happens. The way the brain really solves this physics problem is by taking a shortcut, one which reliably works.

It gets a bit complicated, and there's still debate about which shortcut is used in real life. As I understand it, the fielder uses a (presumably unconscious) heuristic, of running so that as they see it, the ball moves in a particular way, within their field of vision. "The fielder watches the flight of the ball; constantly adjusting her position in response to what she sees. If it appears to be accelerating upward, she moves back. If it seems to be accelerating downward, she moves forward."

This trick would be much easier to show someone, with a bat and ball out on a field, than to describe in words. Although the best outfielder in the world hasn't ever been either shown this trick, or had it described to them.

This all captures a central idea in embodied cognition - that just because a real-life problem can be solved by a computer, that doesn't mean that animal or human brains actually do so, like a computer. Computation can be 'offloaded' onto the body and the environment. We didn't start off as a modern computer in a sea cucumber body; the brain and the body evolved concurrently, and interdependently.

*Research actually refers to 'fly balls' in baseball, thanks a lot American cultural imperialism.

*****

Embodied simulation as a theory of language

Where does language come from? More specifically, what is the mechanism by which the human brain turns infants with no language into fluent speakers? You might think we'd have a clear answer to that by now, but far from it.

When behaviourist theory dominated, in the first half of the 20th century, it tried to explain language acquisition as merely the reinforced association of sound patterns (words) with objects and actions, handed down from one generation to the next, one behavioural training session at a time. Too easy.

Cognitive science, and Noam Chomsky especially, came along in the 1960s and ripped that all to shreds. Chomsky then, and Steven Pinker now, propose an innate, hardwired program for language, which the human brain has evolved - a 'generative grammar'. A really innate program - in the genes. Chomsky talked about 'the poverty of the stimulus', arguing that young kids can't possibly hear enough instances of words and grammar to work out the structure of a human language (the feat can indeed seem miraculous, if you think about it).

But then by this model, what base language does our genes program, which a child then turns into one of the thousands of human languages around? What is the language of thought itself? A 1975 book on the topic by a guy called Jerry Fodor was influential. Pinker later coined the term 'mentalese' for this postulated language of thought.

Some thinkers and researchers today don't believe in this model, however. And for what it's worth, neither do I.

One alternative theory of the neural/cognitive foundation of language is the 'embodied simulation' theory. By this theory, the way all our brains have attached a particular meaning to a word or utterance, is by associating that utterance with a particular sensorimotor experience. A particular sensorimotor experience which we have actually experienced in real life, once or repeatedly.

This theory is closely tied to the 'conceptual metaphor' theory, which argues that most thought (not just language) arises as metaphor; a literal meaning being (largely unconsciously) cognitively recycled for an abstract meaning. The conceptual metaphor that 'anger is fire' is an example - we might consider a remark to be 'inflammatory', without any affectation of poetry. It's argued that these connections to and from conceptual metaphors are so direct that they underlie common fallacies and biases. We might conceptualise linguistic expressions as 'containers for meaning', but that conceptual metaphor might leave us blind to the role which context plays, in determining the meaning of a particular expression in real life.

This can all get very complex and very philosophical, quite quickly. Modern researchers in language embodied simulation have, however, done some pretty cool experiments supporting their claims, which I won't go into here.

To get slightly off topic, I myself suspect that, as is the usually the case in child development, child language acquisition is neither predominantly innate (genes/brain), nor predominantly determined by the environment/body. It's not just some 50-50 split either. It's the product of millions of agent-environment interactions, occurring in real time.

Or, as someone more clever than me has put it (I forget who), the cause of child language acquisition is evenly split between 100% genetic cause and 100% environmental cause.

[The discussion above excludes many other important aspects and frameworks of understanding human language. Those other frameworks, looking at the question from a different angle (such as the social aspect, set out by Jerome Bruner and Michael Tomasello), are of course equally important in explaining language.]

Thursday, 23 January 2020

The zebrafish model of autism

I have to open this post by assuring readers that this is NOT SATIRE.

Go to google scholar ( https://scholar.google.com.au/ ) and type in 'zebrafish', and 'autism'. About 13,000 results? That's what I got.

A couple of these papers were published in 2014, both in journals with 'impact factors' above 10. One paper has been cited 455 times already; the other 305 times.

Titles:
"Zebrafish as an emerging model for studying complex brain disorders", and
"Zebrafish models for translational neuroscience research: from tank to bedside".

This is a very real, established, and publicly funded research program. It encompasses a "Zebrafish model of autism".

About now, you might be a bit perplexed. Perhaps it's all actually about studying genes and proteins, in which anomalies wreak non-specific havoc on all sorts of basic cellular function, in affected primates as much as in fish? That might make sense. But no.

"For brain disorders such as autism ... the use of zebrafish to model social deficits is based on their rich social behaviors. For example, zebrafish are highly social animals and swim in shoals, the disruption of which by various environmental, pharmacological or genetic factor can be easily assessed." See figure below.

The suggestion isn't just that a molecular abnormality which generally causes problems for zebrafish might also cause problems for humans. The suggestion is that the same molecular abnormality which makes zebrafish behave unsociably might also make some humans autistic.

*****

The zebrafish has become popular in laboratories around the world, because "it ticks just about every box on a wish-list of attributes that scientists seek when trying to model human diseases". [ https://www.theguardian.com/science/2013/sep/15/zebrafish-human-genes-project ]

Documented benefits of the zebrafish, to the labs and scientists using it, comprise a long list:

The zebrafish is small and robust.
They are cheaper to maintain than mice.
Break of daylight triggers mating in zebrafish (many other fish only lay eggs in the dark).
Zebrafish produce hundreds of offspring at weekly intervals providing scientists with an ample supply of embryos to study.
They grow at an extremely fast rate, developing as much in a day as a human embryo develops in one month.
Zebrafish embryos are nearly transparent which allows researchers to easily examine the development of internal structures. Every blood vessel in a living zebrafish embryo can be seen using just a low-power microscope.
As zebrafish eggs are fertilised and develop outside the mother’s body it is an ideal model organism for studying early development.
Zebrafish have a similar genetic structure to humans. They share 70 per cent of genes with us. [ https://www.yourgenome.org/facts/why-use-the-zebrafish-in-research ]

The zebrafish model of autism is poor science, sure. Some zebrafish research might be proper science. Some zebrafish research might, indirectly and over the long term, help contribute to greater understanding of some causes of autism.

But the zebrafish model of autism, as it's currently conceived, isn't a great model.

Is that all there is to this? Shall we sit back and chortle at the millions of research dollars being hoovered up by the International Zebrafish Neuroscience Research Consortium (I keep telling you, no - not satire). Or could there be bigger fish to fry (my apologies)?

I think there could be. I think something else could be going on here as well.

In the absence of competitors, the zebrafish model of autism has come to occupy a small corner of a very big space - the autism explanatory gap. What causes autism? What are the pathophysiological steps which conclude in an individual's autistic makeup? What is autism anyway?

No one claims to have a good answer to these questions. Baron-Cohen's 1985 'theory of mind' theory is probably still the best-known attempt at an explanation. I don't find it a satisfactory explanation, at all. Its pre-eminence seems to be fading in recent years. Baron-Cohen published his theory 41 years after Kanner's landmark paper. 35 years have now passed since Baron-Cohen's publication.

Back to the zebrafish. There's an old joke which captures the logic of the zebrafish model of autism, better than I could describe (or satirise) it. The nearly transparent zebrafish embryo adds allegorical power:

A policeman sees a drunk man searching for something under a streetlight and asks what the drunk has lost. He says he lost his keys and they both look under the streetlight together. After a few minutes the policeman asks if he is sure he lost them here, and the drunk replies, no, he lost them in the park. The policeman asks why he is searching here, and the drunk replies, "this is where the light is".

Papers mentioned:

Kalueff A et al., "Zebrafish as an emerging model for studying complex brain disorders" (2014) 35 Trends in Pharmacological Sciences 63.  https://www.sciencedirect.com/science/article/abs/pii/S0165614713002290

Stewart AM et al., "Zebrafish models for translational neuroscience research: from tank to bedside" (2014) 37 Trends in Neurosciences 264.  https://www.sciencedirect.com/science/article/pii/S0166223614000277

Baron-Cohen, S et al., "Does the autistic child have a theory of mind?" (1985) 21 Cognition 37.  http://pages.uoregon.edu/eherman/teaching/texts/Baron-Cohen%20Leslie%20Frith%20Does%20the%20Autistic%20Child%20Have%20a%20Theory%20of%20Mind.pdf

Kanner, L, "Early infantile autism" (1944) 25 Journal of Pediatrics 211.  https://www.jpeds.com/article/S0022-3476(44)80156-1/abstract

Monday, 13 January 2020

Why don't we just measure attention?

A common topic in paediatric consultations is a kid's problem with 'attention' (or concentration, or focus etc.).

It's worth thinking about what people really mean when they say that a child has an 'attention problem'.

Attention deficit hyperactivity disorder - 'ADHD' - remains controversial. Especially the associated use of particular medications in children.

Wouldn't it all be a lot easier if we could just measure a child's attention control ability? We can measure intelligence, after all. Sure, the concept of IQ still fuels hot arguments, especially around race, poverty, and opportunity, and rightly so. But 'IQ' does mean something. If a young child is found to have an IQ of 50, no one would claim that their life prospects are just as bright as any other child's, or that they won't benefit from extra support.

Similarly, 'paying attention to' something, or when something 'comes to your attention', these are real concepts. We all have a similar, plain English, understanding of these terms.

So why hasn't anyone developed an operationalised instrument to measure attention, as has been done for intelligence? Then kids who score under, say, 70 for their 'Attention Quotient' could receive a formal, objective ADHD diagnosis (and a fast-track to an authority script for stimulant medication). Come on, brainiac neuropsychologists, what are you waiting for?

It wouldn't work, of course. I doubt there'll ever be a standardised operationalised instrument to measure someone's attention control ability, in real life. Too may confounders cause too much disruption, in any situation. Anyone's attention control ability is influenced, moment to moment, by a vast range of human, often physical factors. Hunger, sleep deprivation, emotional state, you name it. And even more profoundly by aspects of higher mental function - motivation, confidence, engagement, novelty, familiarity.

(So how then is ADHD 'diagnosed'? That's a question for another post. In which the widespread practice of calling a parent/teacher questionnaire an 'ADHD assessment' will be examined.)

Also, there are different facets to one's attention control ability. There are older concepts like 'sustained attention' (how long I can keep my attention on one object), and 'selective attention'. And related concepts like 'attentional set-shifting'. Attention control ability might depend as much on one's ability to inhibit task-irrelevant motor responses, or to disengage from information recently attended to but no longer useful, as it does on one's raw attentional stamina.

Modern neuropsychology has settled on 'executive attention', as a (quite complex and technical) catch-all term.

For one coherent, interesting, model of this, see Tiego J et al., "A hierarchical model of inhibitory control" (2018) 9 Frontiers in Psychology 1339.
https://www.frontiersin.org/articles/10.3389/fpsyg.2018.01339/full

*****

One more feature of attention control - it's hard to think or talk about the concept, without becoming immediately entangled in questions which seem slippery, or worse still, philosophical.

A kid has more trouble than his peers, in attending to the teacher's classroom instruction. He really wants to be good. Each morning on the way to school, he consciously tells himself that he must listen to the teacher today, and not get in trouble. But in the moment, in the classroom, he can't help but be distracted by something shiny and new. He can't resist the temptation to loudly crack the joke which an opportunity desperately calls for.

This is how I myself might describe events in the life of a 'classic ADHD' kid, in these terms. But if someone were to press me, on just what I mean by 'wants to be good', 'consciously tells himself', 'can't help but be distracted', etc., I'd find it difficult to elaborate further. There aren't many explanatory stops between these terms, and the timeless philosophical mysteries of 'consciousness' and 'free will'.

Further reading:

Engle RW, "Working Memory and Executive Attention: A revisit" (2018) 13 Perspectives on Psychological Science 190. http://englelab.gatech.edu/articles/2018/Engle%20POP%20revisit%20of%202002%20paper.pdf

Heisler JM et al., "The attentional set shifting task: a measure of cognitive flexibility in mice" (2015) 96 Journal of Visualized Experiments e51944. https://www.jove.com/video/51944/the-attentional-set-shifting-task-measure-cognitive-flexibility
This is one of my favourite research papers ever.

Tuesday, 7 January 2020

Child “development” - what do we mean?

When helping a young kid learn to use a pencil, it’s worth looking at their feet. Why?

We use our hands best when our whole body is stable. Often, when a kid sits on a chair at a table, their legs don’t reach the ground. This means their trunk isn't stable, so it’s harder for them to learn to use a pencil. They’ll develop pencil skills quicker with their feet flat on the floor.

Some kids (e.g. those with cerebral palsy) need help to get into a stable sitting position, no matter what size chair they’re in. So, a kid’s mastery of pencil skills sometimes needs a good physio’s help, as much as a good OT’s.

This is a developmental framework - thinking about a child’s progress in terms of the abilities they’ve mastered, they’re currently mastering, or are yet to start on.

A deeper understanding of these abilities, in clinical practice, is the understanding of how these various abilities ‘interact’ with each other. One ability might be important or essential, for the subsequent development of other abilities. This can be obvious, e.g. an infant has to practice babbling, creating lots of different word sounds, before they’ll say their first word. But sometimes the connection isn’t as obvious, as in the feet/pencil example above.

The connections between abilities can be more complex. The development of one ability might need several separate, more basic abilities lined up first, e.g. for a kid to write a word with a pencil, they need fine motor ability with their hands, and early cognitive symbolic understanding of letters and words.

***

A little while ago I was reminded of the power of a developmental framework of understanding. I came across a description of an ability which we all take for granted - understanding the communicative intention of others:
“…Communicative intentions are a special type of intention in which an individual intends something not just toward an inert object but toward the intention states of someone else. Consequently, when an adult addresses an utterance to an infant too young to comprehend intentions, from the infant's point of view the adult is just making noise (for whatever reason). Infants this young may on occasion learn to associate one of these noises with a perceptual event in much the same way a household pet may understand that the sound ‘dinner’ heralds the arrival of food. But this is not language. Sounds become language for young children when and only when they understand that the adult is making that sound with the intention that they attend to something.”

Sometimes I see a very young kid, aged say 1 to 3yo, whose parents are worried because the kid “doesn’t listen”, “won’t follow instructions”, “doesn’t ask for things” or “doesn’t tell us anything”. Even though it seems that the kid is learning some words, and what each word means. In the consultation room, the same kid might wander around the room inspecting things and trying to manipulate them, largely oblivious to myself, and even their parents. The child might glance at me to check me out, and might even smile if I pull a silly face. But they don’t say anything to me, or listen to anything I try to tell them.

Reading about this concept, “understanding the communicative intentions of others”, helped clarify for me what might really be going on. And in a way which can be conveyed to others, especially the child's parents, desperate for meaningful explanations. It can be hard for a parent to understand how a child who knows the word “milk”, who can recognise a glass of milk in a picture book and say "milk" in that situation, doesn't just come out and say “milk” when they want a glass of milk. Instead of taking the parent to the fridge by the hand, or having a meltdown when their unspoken desire isn’t understood.

***

There are other actions and behaviours which come up a lot in clinical practice, but which aren’t really developmental abilities. One example is ‘making eye contact’ (which usually means ‘making eye contact in the culturally accepted, socially appropriate way’). Most young kids will make eye contact with adults they meet, especially when it’s clear that their parents like and trust this new grownup. But some kids don’t, especially not with unfamiliar adults. And their failure to is noticed. The child might be admonished by their parents, that they shouldn’t be so rude to Auntie Beryl. Or the child might end up in a paediatrician’s consulting room.

This detail about the child, that they rarely or never make eye contact, is one of myriad important details which a clinician must try to make sense of, if they’re going to help the kid and family.

But it’s not helpful to think of making eye contact as a developmental ability. It doesn’t fit into that framework well - there's nothing wrong with the kid's visual acuity, or extraocular muscle control. A different framework(s) will make better sense of what’s really going on for the kid.

The quote is from Tomasello M, "Constructing a Language" (2003) Harvard University Press.