between autism and schizophrenia


Neurodiversity and a fine-tuned spectrum

Autistic means Rainman and the schizophrenic are those in the madhouse. While many would immediately agree that this view is somewhat reduced, even the professional view of »autistic« or »schizophrenic« is inevitably oversimplified. Although the diagnosis is more and more differentiating DSM-5, it is still a little bit like assigning »tall« or »short«, instead of giving the body height. It was realized that these conditions span a spectrum1 2, ranging from the neurotypical to the severely impaired. Naturally, most attention is directed towards the severe cases at the end of the spectrum, where people are most seriously affected. However, the presence of such a spectrum implies that there exist a large number of people who may not be affected in a pathological sense, but who still don't feel quite normal. Like someone not strikingly tall, but who still doesn't sit all too comfortable in a long flight. An increased awareness for the neurodiversity, which is the result of such a spectrum, and the particular preferences of every individual is therefore an important development. It is opposed to the classical binary view of »healthy« or »ill«.

It has been proposed, that the autism spectrum and the schizophrenia spectrum extend to two opposite sides from the neurotypical, that they are diametrical opposites3. In other words, while you cannot be purely autistic and schizophrenic at the same time, everyone is somewhere in between. Most are quite in the center, some tend to the »autistish« side, some to the »schizophrenish« side. In a simplest possible model, the population would be normally distributed between the poles. This diametrical spectrum is not the only model. To be honest, the current attention is even more on »overlap« and »sharing« - my prediction is that this fashion (yes, there are fashions in science) will pass, leave insights (some will be discussed later), but that a core diametrical spectrum will remain. And that this diametrical model tells us an awful lot about neurodevelopmental disorders.

The assumption of a spectrum for autism and schizophrenia makes one conclusion almost inescapable: that the spectrum is not purely a consequence of »something-went-wrong« and illness, but that it is evolutionary »intentional« and fine-tuned4. Just as the body height is5. That is not to belittle the illness at the extremes of the spectrum. Indeed, not all identified genetic links to autism or schizophrenia are evolutionary stabilized: some of the genetic risk factors are rare mutations, so the »illness« view is not entirely inappropriate6 7 8 B. Crespi. The diagnosis of these disorders also coalesces with more general cognitive disability. For example some newly arising (not inherited) genetic mutations increase at the same time the risk for autism, intellectual disability and schizophrenia9. Even though there is illness in these neurodevelopmental disorders, it gets more and more evident that most of the genetic risk comes from common variants, that is variants evolution has selected for10 11. In other words, because the larger part of the genetic base for autism or schizophrenia is stabilized by evolution, there has to be a reason for it. Many of the impairments on the autism-schizophrenia spectrum might be, from an evolutionary perspective, a necessary price to pay. Similar to the relationship of the evolutionary variation of body height with growth defects5.

This is to all who don't feel quite at home in our world: the world needs you!

Up to a few percent of people are / could be diagnosed with the neurodevelopmental disorder autism. The simple binary view of autistic and neurotypical, however, is insufficient, as there exists a broad spectrum in severity of the symptoms. As a first step, milder forms were defined. To find definite diagnostic criteria to distinguish between »autistic« or »not autistic« proved to be difficult. It seems that more people are affected by milder forms and that the numbers increase, the more detailed the view is.
It has been realized, that a continuous spectrum describes the reality better, with a decreasing number of affected people the higher the severity of the symptoms gets. The problem is: such a one-sided spectrum for a heritable characteristic simply does not exist! The factors which are varied to make people phenotypically more autistic, can also be varied in the other direction. What do we find on the other side? Only neurotypical?
A diametrical spectrum with schizophrenia on the one end and autism on the other end has been proposed. Evolution is selecting for a dynamic optimum between the two poles. Most attention is typically directed to the extremes of this spectrum. However, one of the most important things to derive from the idea of such a spectrum is the in-betweens: the large number of people who are neither »ill« nor perceive themselves as »normal«. Who are »schizophrenish« or »autistish«.

Evolutionary origin

But where is the evolutionary purpose of such a spectrum to be found? What is varied in the spectrum? It must be significant, when it is paid with some percent of the population being so severely impaired. Special talents which can be found on the spectrum may give some hints. Prominent examples are the savant skills, found on the autism spectrum12, or extraordinary creativity, found on the schizophrenia spectrum13 14 15.

Before we proceed, at quick reminder: we cannot find a monolithic »explanation« or »cause« for autism or schizophrenia. It has been realized long ago16 and was also proclaimed as explicitly as it can be, as in: »Time to give up on a single explanation for autism«17, that these disorders are based on a complex interplay of many factors and that a single cause is not to be found. Even so, to this day, the temptation for a one-dimensional causal chain from cause to symptom seems to be almost too big to resist. This is, because in our intellectual evolution we just arrived at a point to acknowledge complexity, but we still love simplicity or complicatedness much more. The question is, if a single cause is not to be found, whether a single, or at least dominant, effect can be assigned to the autism-schizophrenia spectrum. After all, psychiatrists seem to be able to recognize enough of a pattern to classify in neurotypical, schizophrenic or autistic.

This effect would obviously be found within higher cognitive functions, because this is where the diagnostic criteria are defined. It must be a fundamental emergent parameter of the complex information processing system human brain, not an effect based on a single genetic or molecular biological cause. Something similar to the IQ. It would be fine-tuned by evolution and span the autism-schizophrenia spectrum. Let us think about the information bandwidth at the highest information integration level in the human brain. On the level of consciousness. It is determined by the steady state balance of bottom-up information aggregation and top-down stabilization at the highest information integration level in the human brain. As we will see, it can be linked very nicely with the symptomatic spectrum of neurodevelopmental disorders.

The definition can happen on the background of the »adaptive resonance theory«, a »cognitive and neural theory of how the brain autonomously learns to categorize, recognize, and predict objects and events in a changing world« 18.

Analogue to the variation of body height, which determines the fitness in a diverse environment from directly and indirectly related perspectives, the fine-tuning of the information bandwidth selects for best fitness for information processing in a diverse and constantly changing environment.

This gives a framework for complex neurodevelopmental disorders, which is based on a meta-phenomenon of information processing and serves as a cue for their evolutionary origin.

The ultimate goal of such a framework is being able to measure states within the neurodiversity and then being able to provide appropriate environments. Sure, this will never be possible everywhere; some things will remain to be prepared for the norm, like the seats in an airplane are adjusted to the average body height. But others can be easily matched to individual preferences. We do have more than one size for clothing, don't we? So, why should eduction or work environments consider only the norm?

The goal: being able to measure and differ.

Information bandwidth of conscious access

Let us first try to get an idea what the information bandwidth of conscious access is. It is the amount of information maintained by resonant states in the highest level in the hierarchy of information integration. Our brain is fed with an immense amount of sensory input. One primary purpose of neural processing is to integrate the large amount of sensory information to a millionth of the original bandwidth and to come to a certain, hopefully advantageous, output Tor Nørretranders. The output steers action by again expanding the information bandwidth to ultimately control every single muscle fiber. Depending on the type of input and action, the processing can be parallelizable well, not so well, or not at all. All non-parallelizable processing needs to be synchronized at the highest integration level, while parallelizable processing can happen independently at lower integration levels. This leads to gradually increasing levels of hierarchy at which reductive input processing switches to expansive output processing. Processing of independent input from a non-varying environment is generally well parallelizable, with a set of input always leading to identical output. An example is processing for reflex motor activity. Other actions, mostly if they need to compete for resources with other actions or which respond to a variable environment, require hierarchical integration over a broad spectrum of input information. This results in a hierarchical processing where the amount of parallelized processing is reduced in each step by reciprocal inhibition or excitation. Upon this process a set of processed information is stabilized in a »winner takes all« manner19. This set of integrated information ultimately enters, or better, establishes consciousness20.

It's about a fine-tuned information bandwith of conscious access.

The brain, as an information processing system, reduces the bandwidth of incoming sensory information by massive information integration. This happens by matching input with learned patterns, which allows prediction and computation of »advantageous« output. A transition from bottom-up information integration to top-down output can occur at different integration levels (white arrows). After the transition, the information bandwidth has to be expanded again, to ultimately control every single muscle fiber. In contrast to most machine computation, the default processing mode is parallel.
Many cognitive functions cannot be based on parallel and independent processing alone. For output which requires a high integration and abstraction level on a global scale, for example mutual exclusive output (like you cannot both turn left and right at the same time), integration of multiple parallel and incoherent streams of information is needed. Therefore, selected information in the highest integration level is maintained for some time in a »winner takes all« manner. The amount of information in this highest hierarchy level is the bandwidth of conscious access, estimated to comprise some tens of Bits per second.
Feedback to lower level brain areas stabilizes these streams of information and inhibits other competing streams. This feedback generates resonant, conscious states as neuronal correlate of consciousness. If the transition from sensory information processing to output happens at this highest integration level we call it »decision making«.
Another function of the brain that depends on high information integration is communication. In order to use narrow communication channels (like language), information has to be highly compressed.
The highly integrated and compressed information at the highest integration level can be fed back into the processing system for formation of long term memory. Similarly, feeding back and »looping« information can lead to further integration with information from memory, a process we call thinking and reasoning.

There are several cognitive functions that build on this wide and hierarchical information integration: communication, memory formation, decision making and reasoning. The bandwidth of communication channels is extremely narrow compared to the bandwidth of information processed in parallel. Therefore extensive reduction is necessary to use the narrow channel of language for relevant information (a prototypal example for information processing which has to compete for resources). The same applies for the formation of (declarative) memory. It is not desirable to memorize raw sensory data. Each memory contributes to the input which has to be processed later. Ultimately this means that if raw information was stored, the neuronal system would need to grow exponentially with the amount of stored information. Since this is not possible, only storage of preprocessed information takes place. Only the information »considered most important« enters consciousness and is forwarded to the memory formation machinery. Mutual exclusive actions also need hierarchical information prioritizing for decision making. When parallel action is not possible, one output has to be singled out as »best« solution. A further usage of this hierarchical processing has developed in form of (logical) reasoning. Conscious information is fed back into a lower hierarchical level to be reprocessed. This is especially useful to train the neuronal network towards a more robust result, if the output is variable.

The key evolutionary challenge is now, that the narrow bandwidth of highest-level information needs to be exactly tuned for the best fitness in diverse situations and in a constantly changing environment. A trade-off between computation-time and -depth needs to be made for best-possible outputs. This trade-off cannot look the same for a CEO in a fast pace business and a theoretical physicist. For an artist or an engineer. It is the ongoing matching of the ideal bandwidth of integrated information and conscious access, achieved by balancing the input bandwidth and stabilization, which is at the core of the autism-schizophrenia spectrum.

Variations in the bandwith of conscious access

All main functions of the highest integration level can be more or less affected in people with autism spectrum disorders. On the autistic spectrum, the bandwidth of conscious access is enlarged. »Sensory overload« is the consequence. Since the bandwidth of the communication channel is fixed (you cannot say two words at the same time), communication capabilities can be affected. The autistic spectrum ranges from atypical language use (Asperger Syndrome) over slightly impaired language skills in High-Functioning-Autism (HFA) to entirely missing communication. Some autistic individuals may not be able to communicate, but are considered savants and for instance can memorize a visual image in a level of detail that is far from what neurotypical people can achieve. As impressive this may be, it comes at a price. The memory formation is shifted from a higher integration level down to lower hierarchical levels and only weakly integrated data is memorized.

As overstraining daily situations can be for people on the autism spectrum, as impressive their ability for logical reasoning under the right conditions can be. The enlarged bandwidth of conscious information can help while working on complex problems where the neurotypical person tends to miss the concealed but important little clue. This can be considered as a strong tendency to convergent versus divergent thinking. The incomplete filtering of sensory input leads to manifold sensory abnormalities autistic people can suffer from. Under-responsivity to salient input can be caused by the flooding of consciousness with information. At the same time, the inability to blind out certain low-merit sensory input can also lead to over-responsivity. The higher the integration level is a certain action requires, the more likely it is to observe impairment in autistic individuals. Other symptoms of ASD like repetitive behavior however, do not fit into this scheme. They can be considered as a compensatory reaction to the constantly overwhelmed mental state in varying environments which is linked to a high level of stress. A less variable input to the neuronal system is enforced by repetitive behavior.

More and more input: towards autism.

Over-increased integration and reduction of input information leads to the symptoms of schizophrenia. The opposite leads to the symptoms of autism.

If the enlargement of the bandwidth of consciousness can lead to autism, what happens if the bandwidth is narrowed? As far as communication is concerned, it can happen that information is communicated which appears unexpected, short sighted and illogical to the conversational partner. This is a typical symptom for schizophrenia. The hierarchical processing tends to get inconsistent and misses important information. This also can impair logical reasoning, since it works on a limited and constantly varying subset of input. Consciousness is lacking a balanced information input and therefore can start to occupy itself with repetitive input loops which lack consistency with reality. Since the content of consciousness is fed back to memory formation psychoses can develop. Before that happens, earlier on the spectrum a kind of »immunity« to constant disturbance from sensory input is granted. An ability to reach out far away from the »typical« information integration develops. Many artists who surprise with their genius are on this part of the spectrum. Their ability to combine colors, tones, words (and all the other things which were inconceivable for others) in a way that concentrates on certain aspects, brings them together in a uncommon and mind bending way. Their intuition may be result of a tendency of neuronal processing which is inverse to the tendency of the autistic spectrum. While the autistish side of the spectrum tends to integrative and convergent processing, the schizophrenish side tends to divergent and creative thinking. This promotes creativity13 14 and favors novelty and flexibility over logical consistency. As far as decision making is concerned, people on the schizophrenic spectrum can appear impulsive and short sighted. Also, the repeated experience of having made wrong decisions because not having considered all important facts can lead to reluctance in decision making.

These considerations lead to the conclusion that autism and schizophrenia should not be seen in a binary discrimination as mental illnesses compared to mental health, but as endpoints of a evolutionary necessary spectrum. The evolutionary cause for a spectrum of neuronal processing between autism and schizophrenia is the balancing act between consistent information integration and fast and creative response to a variable environment.

A high local connectivity increases the size of neuronal correlates of consciousness (NCC) by increased excitation, whereas an increased long-range connectivity – most importantly with the thalamus and the basal ganglia – increases the stabilization time of NCC and the bias towards top-down versus bottom-up attention. High local connectivity and low inhibitory connectivity to the basal ganglia leads to a large population of neurons in the NCC and a low stabilization time and therefore to a high bandwidth of conscious information.

More and more information reduction: towards schizophrenia.

Twodimensional spectrum - between activation and stabilization

schizophrenia positive negative21 22 23 autism social deficits, communication and repetitive behaviour DSM-5, genetic diff 24 Asperger syndrome 25 26 ASP, HFA anxiety 27

Dimension one
↑ ↑ ↑
  • Striatal dopamine hyper-function 28
  • Overactive »default network«29
  • Thalamocortical dysconnectivity 30
↓ ↓ ↓
  • Rare CNV: 22q11.2 duplication31
  • Long range under-connectivity 32
  • Impaired thalamocortical connectivity 33
Two-dimensional spectrum of neurodevelopmental disorders spanned by the amount of cortico-cortical activation in one dimension and thalamocortical/subcortical stabilization in the other dimension. The neurotypical is found where input and stabilization are ideally balanced. NT: neurotypical. A: autism. S: schizophrenia. AS: asperger syndrome. HFA: high functioning autism. SPD: schizoid personality disorder. STPD: schizotypal personality disorder. ADHD: attention deficit hyperactivity disorder. ADD: attention deficit disorder. The position on the spectrum also marks the susceptibility for certain mood disorders (blue dashed lines). AMD: anxiety mood disorder. DMD: depression mood disorder. BPMD: bipolar mood disorder. BLMD: borderline mood disorder.
Dimension two
→ → →
  • Local over-connectivity 32
  • Hyper-functional neural microcircuits 34
  • Brain overgrowth at age 1-5, particularly in PFC 35 36
  • Under-specific amygdala response 37
  • Shank3 loss-of-function 38
← ← ←
  • Shank3 duplication 39
Dimension three
↑ ↑ ↑
↓ ↓ ↓

3D 2D

Are two dimensions sufficient? Evidence for common genetic and environmental factors which increase the risk for both autism and schizophrenia is accumulating. These facors can be summarized in a third »susceptibility« dimension. An important example is, when the brain looses ability for auto-balancing bottom-up activation and top-down stabilization. If this ability is impaired, depending on the predisposition, the risk for both neurodevelopmental disorders is increased. NT: neurotypical. A: autism. S: schizophrenia.

To be continued...

More than half a century of autism research allow the conclusion that there is no single genetic or cognitive cause to be found. On the operative level, however, an abnormal bandwidth of consciousness appears as comprehensive explanation. In the neuronal representation, this bandwidth may be governed by the establishment and stability of neuronal coalitions formed in the frontal and temporal cortex, which form the neuronal correlates of consciousness. The correlation of neurodevelopmental disorders and mental disorders with consciousness can advance understanding and research of consciousness fundamentally. All collected knowledge of genetics and neurobiology of these disorders can give insight to the basic principles of consciousness.

Latent inhibition - extraversion

Neural processing pathways are established by repeated successful loops from sensory input to (motor) output. Processing of information with high entropy, that is, information processing which cannot maintain a certain output for certain isolated subsets of information, require overall processing. Ultimately, this means each neuron has to be interconnected with every other neuron. This is, of course, not possible literally. The interconnection is achieved indirectly by 1) a shift from explicit representation to implicit representation 2) hierarchical processing. The flood of information is integrated to implicit, abstract concepts as much as possible or necessary, which are subsequently offset against each other. This finally forms the highest hierarchical level. The components and the bandwidth of this highest level determine the “mental state” of the neural system. If all common, human ingredients contribute in this mental state in an ordinary relative proportion, this mental state can be called consciousness. It is certainly not a single, exclusive feature which generates a conscious mental state, but a certain combination of several core features. A prominent example for these ingredients is a theory of mind10 which is considered necessary for a conscious mental state. The definition and inherent nature of consciousness remains a debate, therefore it is more reasonable to define consciousness, as used here, with this highest hierarchical level of information processing in a neuronal system. The bandwidth of which then spans the spectrum between autistic, neurotypical and schizophrenic individuals. This saves the debate whether severe autistic or schizophrenic people can be called conscious or not – they certainly are not in a classic sense. But how is this hierarchical information processing realized on neuronal level? Cortical-thalamic feedback loops appear to be an important monitor of motor instructions- the thalamic relay function might be responsible for blocking parallelizable processing for higher hierarchical levels. What happens to signals, which cannot establish subconscious motor instructions in thalamocortical pathways? Observations of neuronal abnormalities of individuals with autism and schizophrenia can help to set up a framework. Relevant observations are: A reduced volume of the prefrontal cortex in schizophrenia and an increased brain volume, at least in the first years in autism11. Local overconnectivity in the prefrontal cortex and long-range underconnectivity in autism12. Dereased fractional anisotropy (FA) in anterior thalamic radiation [12,13] and uncinate fasciculus [14]13 Increased FA in autism14 A decreased subcortical tonic dopamine level, with subsequent oversensitivity to phasic dopamine release. Can be caused by reduced prefrontal activation (glutamergic) of the striatum.

deletion schizophrenia, duplication autism

genetics difficult to investigate GWAS

5Lateralization of cognitive processing and bandwidth of consciousness.

amount of integrated information ADHD/ autism 40 ADHD schizo, negative symptoms 41

1Baron-Cohen, S.; Wheelwright, S.; Skinner, R.; Martin, J.; Clubley, E. The tism-Spectrum Quotient (AQ): Evidence from Asperger Syndrome/High-Functioning Autism, Males and Females, Scientists and MathematiciansJ. Autism Dev. Disord. 2001  pp. 5–17 
2Sbrana, a.; Dell’Osso, L.; Benvenuti, A.; Rucci, P.; Cassano, P.; Banti, S.; Gonnelli, C.; Doria, M. R.; Ravani, L.; Spagnolli, S.; et al. The psychotic spectrum: Validity and reliability of the Structured Clinical Interview for the Psychotic SpectrumSchizophr. Res. 2005  pp. 375–387 
3Crespi, B.; Stead, P.; Elliot, M. Evolution in health and medicine Sackler colloquium: Comparative genomics of autism and schizophrenia.Proc. Natl. Acad. Sci. U. S. A. 2010  pp. 1736–1741 
4Gualtieri, C. T. Autism and Schizophrenia Are Disorders of EvolvabilityOpen J. Med. Psychol. 2014  pp. 161–183 
5Lango Allen, H.; Estrada, K.; Lettre, G.; Berndt, S. I.; Weedon, M. N.; Rivadeneira, F.; Willer, C. J.; Jackson, A. U.; Vedantam, S.; Raychaudhuri, S.; et al. Hundreds of variants clustered in genomic loci and biological pathways affect human height.Nature 2010  pp. 832–838 
6Malhotra, D.; Sebat, J. CNVs: harbingers of a rare variant revolution in psychiatric genetics.Cell 2012  pp. 1223–1241 
7Pinto, D.; Pagnamenta, A. T.; Klei, L.; Anney, R.; Merico, D.; Regan, R.; Conroy, J.; Magalhaes, T. R.; Correia, C.; Abrahams, B. S.; et al. Functional impact of global rare copy number variation in autism spectrum disordersNature 2010  pp. 368–372 
8Purcell, S. M.; Moran, J. L.; Fromer, M.; Ruderfer, D.; Solovieff, N.; Roussos, P.; O’Dushlaine, C.; Chambert, K.; Bergen, S. E.; Kähler, A.; et al. A polygenic burden of rare disruptive mutations in schizophrenia.Nature 2014  pp. 185–190 
9Iossifov, I.; O’Roak, B. J.; Sanders, S. J.; Ronemus, M.; Krumm, N.; Levy, D.; Stessman, H. a.; Witherspoon, K. T.; Vives, L.; Patterson, K. E.; et al. The contribution of de novo coding mutations to autism spectrum disorderNature 2014  pp. 216–221 
10Gaugler, T.; Klei, L.; Sanders, S. J.; Bodea, C. a; Goldberg, A. P.; Lee, A. B.; Mahajan, M.; Manaa, D.; Pawitan, Y.; Reichert, J.; et al. Most genetic risk for autism resides with common variation.Nat. Genet. 2014  pp. 881–885 
11Ripke, S.; Neale, B. M.; Corvin, A.; Walters, J. T. R.; Farh, K.-H.; Holmans, P. a.; Lee, P.; Bulik-Sullivan, B.; Collier, D. a.; Huang, H.; et al. Biological insights from 108 schizophrenia-associated genetic lociNature 2014  pp. 421–427 
12Ruthsatz, J.; Urbach, J. B. Child prodigy: A novel cognitive profile places elevated general intelligence, exceptional working memory and attention to detail at the root of prodigiousnessIntelligence 2012  pp. 419–426 
13Jung, R. E.; Grazioplene, R.; Caprihan, A.; Chavez, R. S.; Haier, R. J. White matter integrity, creativity, and psychopathology: disentangling constructs with diffusion tensor imaging.PLoS One 2010  p. e9818 
14de Manzano, Ö.; Cervenka, S.; Karabanov, A.; Farde, L.; Ullén, F. Thinking Outside a Less Intact Box: Thalamic Dopamine D2 Receptor Densities Are Negatively Related to Psychometric Creativity in Healthy IndividualsPLoS One 2010  p. e10670+ 
15Kaufman, S. B.; Paul, E. S. Creativity and schizophrenia spectrum disorders across the arts and sciencesFront. Psychol. 2014  pp. 3–6 
16Petronis, A. The origin of schizophrenia: Genetic thesis, epigenetic antithesis, and resolving synthesisBiol. Psychiatry 2004  pp. 965–970 
17Happé, F.; Ronald, A.; Plomin, R. Time to give up on a single explanation for autism.Nat. Neurosci. 2006  pp. 1218–1220 
18Grossberg, S. Adaptive Resonance Theory: How a brain learns to consciously attend, learn, and recognize a changing worldNeural Networks 2013  pp. 1–47 
19Crick, F.; Koch, C. Towards a neurobiological theory of consciousnessSemin. Neurosci. 1990  pp. 263–275 
20Massimini, M.; Boly, M.; Casali, A.; Rosanova, M.; Tononi, G. A perturbational approach for evaluating the brain’s capacity for consciousnessProg. Brain Res. 2009  pp. 201–214 
21Andreasen, N. C.; Olsen, S. Negative v positive schizophrenia. Definition and validation.Arch. Gen. Psychiatry 1982  pp. 789–794 
22Andreasen, N. C. Positive vs. negative schizophrenia: a critical evaluation.Schizophr. Bull. 1985  pp. 380–389 
23Kay, S. R.; Opler, L. A.; Lindenmayer, J. P. The Positive and Negative Syndrome Scale (PANSS): Rationale and standardisationBr. J. Psychiatry 1989  pp. 59–65 
24Happé, F.; Ronald, A. The “fractionable autism triad”: a review of evidence from behavioural, genetic, cognitive and neural research.Neuropsychol. Rev. 2008  pp. 287–304 
25Attwood, T. Is there a difference between Asperger’s syndrome and High Functioning Autism ? 2003   
26Ozonoff, S.; South, M.; Miller, J. N. DSM-IV-Defined Asperger Syndrome: Cognitive, Behavioral and Early History Differentiation from High-Functioning AutismAutism 2000  pp. 29–46 
27Langdon, P. E.; Murphy, G. H.; Wilson, E.; Shepstone, L.; Fowler, D.; Heavens, D.; Malovic, A.; Russell, A. Asperger syndrome and anxiety disorders (PAsSA) treatment trial: a study protocol of a pilot, multicentre, single-blind, randomised crossover trial of group cognitive behavioural therapy.BMJ Open 2013  pp. 1–6 
28Howes, O. D.; Kapur, S. The dopamine hypothesis of schizophrenia: version III--the final common pathway.Schizophr. Bull. 2009  pp. 549–562 
29Buckner, R. L.; Andrews-Hanna, J. R.; Schacter, D. L. The brain’s default network: anatomy, function, and relevance to disease.Ann. N. Y. Acad. Sci. 2008  pp. 1–38 
30Woodward, N. D.; Karbasforoushan, H.; Heckers, S. Thalamocortical dysconnectivity in schizophreniaAm. J. Psychiatry 2012  pp. 1092–1099 
31Rees, E.; Kirov, G.; Sanders, A.; Walters, J. T. R.; Chambert, K. D.; Shi, J.; Szatkiewicz, J.; O’Dushlaine, C.; Richards, a L.; Green, E. K.; et al. Evidence that duplications of 22q11.2 protect against schizophrenia.Mol. Psychiatry 2014  pp. 37–40 
32Courchesne, E.; Pierce, K. Why the frontal cortex in autism might be talking only to itself: local over-connectivity but long-distance disconnection.Curr. Opin. Neurobiol. 2005  pp. 225–230 
33Nair, A.; Treiber, J. M.; Shukla, D. K.; Shih, P.; Müller, R. A. Impaired thalamocortical connectivity in autism spectrum disorder: A study of functional and anatomical connectivityBrain 2013  pp. 1942–1955 
34Markram, K.; Markram, H. The intense world theory - a unifying theory of the neurobiology of autism.Front. Hum. Neurosci. 2010  p. 224 
35Redcay, E.; Courchesne, E. When is the brain enlarged in autism? A meta-analysis of all brain size reports.Biol. Psychiatry 2005  pp. 1–9 
36Courchesne, E.; Mouton, P.; Calhoun, M.; Semendeferi, K.; Ahrens-Barbeau, C.; MJ, Hallet, M.; Barnes, C.; Pierce, K. Neuron Number and Size in Prefrontal Cortex of Children With AutismYearb. Pediatr. 2013  pp. 395–397 
37Grossberg, S.; Seidman, D. Neural dynamics of autistic behaviors: cognitive, emotional, and timing substrates.Psychol. Rev. 2006  pp. 483–525 
38Boccuto, L.; Lauri, M.; Sarasua, S. M.; Skinner, C. D.; Buccella, D.; Dwivedi, A.; Orteschi, D.; Collins, J. S.; Zollino, M.; Visconti, P.; et al. Prevalence of SHANK3 variants in patients with different subtypes of autism spectrum disordersEur. J. Hum. Genet. 2012  pp. 310–316 
39Han, K.; Holder, J. L.; Schaaf, C. P.; Lu, H.; Chen, H.; Kang, H.; Tang, J.; Wu, Z.; Hao, S.; Cheung, S. W.; et al. SHANK3 overexpression causes manic-like behaviour with unique pharmacogenetic properties.Nature 2013  pp. 72–77 
40Polderman, T. J. C.; Hoekstra, R. a; Posthuma, D.; Larsson, H. The co-occurrence of autistic and ADHD dimensions in adults: an etiological study in 17 770 twinsTransl. Psychiatry 2014  p. e435 
41Donev, R.; Gantert, D.; Alawam, K.; Edworthy, A.; Hässler, F.; Meyer-Lindenberg, A.; Dressing, H.; Thome, J. Comorbidity of schizophrenia and adult attention-deficit hyperactivity disorder.World J. Biol. Psychiatry 2011  pp. 52–56