In the early 1960’s Roger Sperry proposed his famous “chemoaffinity theory” to explain how neural connectivity arises. This was based on observations of remarkable specificity in the projections of nerves regenerating from the eye of frogs to their targets in the brain. His first version of this theory proposed that each neuron found its target by expression of matching labels on their respective surfaces. He quickly realised, however, that with ~200,000 neurons in the retina, the genome was not large enough to encode separate connectivity molecules for each one. This led him to the insight that a regular array of connections of one field of neurons (like the retina) across a target field (the optic tectum in this case) could be readily achieved by gradients of only one or a few molecules.

The molecules in question, Ephrins and Eph receptors, were discovered thirty-some years later. They are now known to control topographic projections of sets of neurons to other sets of neurons across many areas of the brain, such that nearest-neighbour relationships are maintained (e.g., neurons next to each other in the retina connect to neurons next to each other in the tectum). In this way, the map of the visual world that is generated in the retina is transmitted intact to its targets. Actually, maintenance of nearest-neighbour topography seems to be a general property of projections between any two areas, even ones that do not obviously map some external property across them.

But the idea of matching labels was not wrong – they do exist and they play a very important part in an earlier step of wiring – finding the correct target region in the first place. This is nicely illustrated by a beautiful paper studying projections of retinal neurons in the mouse, which implicates proteins in the Cadherin family in this process.

In the retina, photoreceptor cells sense light and transmit this information, through a couple of relays, to retinal ganglion cells (RGCs). These are the cells that send their projections out of the retina, through the optic nerve, to the brain. But the tectum is not the only target of these neurons. There are, in fact, at least 20 different types of RGCs with distinct functions that project from the retina to various parts of the brain.

In mammals, “seeing” is mediated by projections to the visual centre of the thalamus, which projects in turn to the primary visual cortex. But conscious vision is only one thing we use our eyes for. The equivalent of the tectum, called the superior colliculus in mammals, is also a target for RGCs, and mediates reflexive eye movements, head turns and shifts of attention. (It might even be responsible for blindsight – subconscious visual responsiveness in consciously blind patients). Other RGCs send messages to regions controlling circadian rhythms (the suprachiasmatic nuclei) or pupillary reflexes (areas of the midbrain called the olivary pretectal nuclei).

These RGCs express a photoresponsive pigment (melanopsin) and respond to light directly. This likely reflects the fact that early eyes contained both ciliated photoreceptors (like current rods and cones) and rhabdomeric photoreceptors (possibly the ancestors of RGCs and other retinal cells).

So how do these various RGCs know which part of the brain to project to? This was the question investigated by Andrew Huberman and colleagues, who looked for inspiration to the fly eye. It had previously been shown that a member of the Cadherin family of proteins was involved in fly photoreceptor axons choosing the right layer to project to in the optic lobe. Cadherins are “homophilic” adhesion molecules – they are expressed on the surface of cells and like to bind to themselves. Two cells expressing the same Cadherin protein will therefore stick to each other. This stickiness may be used as a signal to make a synaptic connection between a neuron and its target.

The protein implicated in flies, N-Cadherin, is widely expressed in mammals and thus unlikely to specify connections to different targets of the retina. But Cadherins comprise a large family of proteins, suggesting that other members might play more specific roles. This turns out to be the case – a screen of these proteins revealed several expressed in distinct regions of the brain receiving inputs from subtypes of RGCs. One in particular, Cadherin-6, is expressed in non-image-forming brain regions that receive retinal inputs – those controlling eye movements and pupillary reflexes, for example. The protein is also expressed in a very discrete subset of RGCs – specifically those that project to the Cadherin-6-expressing targets in the brain.

The obvious hypothesis was that this matching protein expression allowed those RGCs to recognise their correct targets by literally sticking to them. To test this, they analysed these projections in mice lacking the Cadherin-6 molecule. Sure enough, the projections to those targets were severely affected – the axons spread out over the general area of the brain but failed to zero in on the specific subregions that they normally targeted.

These results illustrate a general principle likely to be repeated using different Cadherins in different RGC subsets and also in other parts of the brain. Indeed, a paper published at the same time shows that Cadherin-9 may play a similar function in the developing hippocampus. In addition, other families of molecules, such as Leucine-Rich Repeat proteins may play a similar role as synaptic matchmakers by promoting homophilic adhesion between neurons and their targets. (Both Cadherins and LRR proteins also have important “heterophilic” interactions with other proteins).

The expansion of these families in vertebrates could conceivably be linked to the greater complexity of the nervous system, which presumably requires more such labels to specify it. But these molecules may be of more than just academic interest in understanding the molecular logic and evolution of the genetic program that specifies brain wiring. Mutations in various members of the Cadherin (and related protocadherin) and LRR gene families have also been implicated in neurodevelopmental disorders, including autism, schizophrenia, Tourette’s syndrome and others. Defining the molecules and mechanisms involved in normal development may thus be crucial to understanding the roots of neurodevelopmental disease.

Osterhout, J., Josten, N., Yamada, J., Pan, F., Wu, S., Nguyen, P., Panagiotakos, G., Inoue, Y., Egusa, S., Volgyi, B., Inoue, T., Bloomfield, S., Barres, B., Berson, D., Feldheim, D., & Huberman, A. (2011). Cadherin-6 Mediates Axon-Target Matching in a Non-Image-Forming Visual Circuit Neuron, 71 (4), 632-639 DOI: 10.1016/j.neuron.2011.07.006

Williams, M., Wilke, S., Daggett, A., Davis, E., Otto, S., Ravi, D., Ripley, B., Bushong, E., Ellisman, M., Klein, G., & Ghosh, A. (2011). Cadherin-9 Regulates Synapse-Specific Differentiation in the Developing Hippocampus Neuron, 71 (4), 640-655 DOI: 10.1016/j.neuron.2011.06.019

However, many perfectly healthy people also regularly hear voices – as many as 1 in 25 according to some studies, and in most cases these experiences are perfectly benign. In fact, we all hear voices “belonging to other people” when we dream – we can converse with these voices, waiting for their responses as if they were derived from external agents. Of course, these percepts are actually generated by the activity of our own brain, but how?

There is good evidence from neuroimaging studies that the same areas that respond to external speech are active when people are having these kinds of auditory hallucinations. In fact, inhibiting such areas using transcranial magnetic stimulation may reduce the occurrence or intensity of heard voices. But why would the networks that normally process speech suddenly start generating outputs by themselves? Why would these outputs be organised in a way that fits speech patterns, as opposed to random noise? And, most importantly, why does this tend to occur in people with schizophrenia? What is it about the pathology of this disorder that makes these circuits malfunction in this specific way?

An interesting approach to try and get answers to these questions has been to model these circuits in artificial neural networks. If you can generate a network that can process speech inputs and find certain conditions under which it begins to spontaneously generate outputs, then you may have an informative model of auditory hallucinations. Using this approach, a couple of studies from several years ago from the group of Ralph Hoffman have found some interesting clues as to what may be going on, at least on an abstract level.

Their approach was to generate an artificial neural network that could process speech inputs. Artificial neural networks are basically sets of mathematical functions modelled in a computer programme. They are designed to simulate the information-processing functions carried out by individual neurons and, more importantly, the computational functions carried out by an interconnected network of such neurons. They are necessarily highly abstract, but they can recapitulate many of the computational functions of biological neural networks. Their strength lies in revealing unexpected emergent properties of such networks.

The particular network in this case consisted of three layers of neurons – an input layer, an output layer, and a “hidden” layer in between – along with connections between these elements (from input to hidden and from hidden to output, but crucially also between neurons within the hidden layer). “Phonetic” inputs were fed into the input layer – these consisted of models of speech sounds constituting grammatical sentences. The job of the output layer was to report what was heard – representing different sounds by patterns of activation of its forty-three neurons. Seems simple, but it’s not. Deciphering speech sounds is actually very difficult as individual phonetic elements can be both ambiguous and variable. Generally, we use our learned knowledge of the regularities of speech and our working memory of what we have just heard to anticipate and interpret the next phonemes we hear – forcing them into recognisable categories. Mimicking this function of our working memory is the job of the hidden layer in the artificial neural network, which is able to represent the prior inputs by the pattern of activity within this layer, providing a context in which to interpret the next inputs.

The important thing about neural networks is they can learn. Like biological networks, this learning is achieved by altering the strengths of connections between pairs of neurons. In response to a set of inputs representing grammatical sentences, the network weights change in such a way that when something similar to a particular phoneme in an appropriate context is heard again, the pattern of activation of neurons representing that phoneme is preferentially activated over other possible combinations.

The network created by these researchers was an able student and readily learned to recognise a variety of words in grammatical contexts. The next thing was to manipulate the parameters of the network in ways that are thought to model what may be happening to biological neuronal networks in schizophrenia.

There are two major hypotheses that were modelled: the first is that networks in schizophrenia are “over-pruned”. This fits with a lot of observations, including neuroimaging data showing reduced connectivity in the brains of people suffering with schizophrenia. It also fits with the age of onset of the florid expression of this disorder, which is usually in the late teens to early twenties. This corresponds to a period of brain maturation characterised by an intense burst of pruning of synapses – the connections between neurons.

In schizophrenia, the network may have fewer synapses to begin with, but not so few that it doesn’t work well. This may however make it vulnerable to this process of maturation, which may reduce its functionality below a critical threshold. Alternatively, the process of synaptic pruning may be overactive in schizophrenia, damaging a previously normal network. (The evidence favours earlier disruptions).

The second model involves differences in the level of dopamine signalling in these circuits. Dopamine is a neuromodulator – it alters how neurons respond to other signals – and is a key component of active perception. It plays a particular role in signalling whether inputs match top-down expectations derived from our learned experience of the world. There is a wealth of evidence implicating dopamine signalling abnormalities in schizophrenia, particularly in active psychosis. Whether these abnormalities are (i) the primary cause of the disease, (ii) a secondary mechanism causing specific symptoms (like psychosis), or (iii) the brain attempting to compensate for other changes is not clear.

Both over-pruning and alterations to dopamine signalling could be modelled in the artificial neural network, with intriguing results. First, a modest amount of pruning, starting with the weakest connections in the network, was found to actually improve the performance of the network in recognising speech sounds. This can be understood as an improvement in the recognition and specificity of the network for sounds which it had previously learned and probably reflects the improvements seen in human language learners, along with the concomitant loss in ability to process or distinguish unfamiliar sounds (like “l” and “r” for Japanese speakers).

However, when the network was pruned beyond a certain level, two interesting things happened. First, its performance got noticeably worse, especially when the phonetic inputs were degraded (i.e., the information was incomplete or ambiguous). This corresponds quite well with another symptom of schizophrenia, especially those who experience auditory hallucinations – sufferers show phonetic processing deficits under challenging conditions, such as a crowded room.

The second effect was even more striking – the network started to hallucinate! It began to produce outputs even in the absence of any inputs (i.e., during “silence”). When not being driven by reliable external sources of information, the network nevertheless settled into a state of activity that represented a word. The reason the output is a word and not just a meaningless pattern of neurons is that the previous learning that the network undergoes means that patterns representing words represent “attractors” – if some random neurons start to fire, the weighted connections representing real words will rapidly come to dominate the overall pattern of activity in the network, resulting in the pattern corresponding to a word.

Modeling alterations in dopamine signalling also produced both a defect in parsing degraded speech inputs and hallucinations. Too much dopamine signalling produced these effects but so did a combination of moderate over-pruning and compensatory reductions in dopamine signalling, highlighting the complex interactions possible.

The conclusion from these simulations is not necessarily that this is exactly how hallucinations emerge. After all, the artificial neural networks are pretty extreme abstractions of real biological networks, which have hundreds of different types of neurons and synaptic connections and which are many orders of magnitude more complex numerically. But these papers do provide aat least a conceptual demonstration of how a circuit designed to process speech sounds can fail in such a specific and apparently bizarre way. They show that auditory hallucinations can be viewed as the outputs of malfunctioning speech-processing circuits.

They also suggest that different types of insult to the system can lead to the same type of malfunction. This is important when considering new genetic data indicating that schizophrenia can be caused by mutations in any of a large number of genes affecting how neural circuits develop. One way that so many different genetic changes could lead to the same effect is if the effect is a natural emergent property of the neural networks involved.

Hoffman, R., & Mcglashan, T. (2001). Book Review: Neural Network Models of Schizophrenia The Neuroscientist, 7 (5), 441-454 DOI: 10.1177/107385840100700513

Hoffman, R., & McGlashan, T. (2006). Using a Speech Perception Neural Network Computer Simulation to Contrast Neuroanatomic versus Neuromodulatory Models of Auditory Hallucinations Pharmacopsychiatry, 39, 54-64 DOI: 10.1055/s-2006-931496

Mirrored from Wiring the Brain