Seed Grants

(Year 1 grant)

Rett Syndrome (RTT) is a devastating progressive autism spectrum disorder frequently associated with loss-of-function mutations in MeCP2, a transcriptional regulator. MeCP2 mutations affect expression of a wide range of transcriptional targets, including multiple signaling proteins such as IGF-1, IGFBP3, PTP1b, and BDNF. Intriguingly, RTT phenotypes can be ameliorated through modulation of signaling networks by supplementation or inhibition of critical signaling components, demonstrating a role for signaling networks in regulating RTT symptoms. While promising, these studies have focused on a few select signaling components based on a priori knowledge or MeCP2 transcriptional targets. We hypothesize that an unbiased comprehensive analysis of signaling network alterations occurring during RTT development and progression will identify multiple additional critical signaling nodes that govern the progression of the disease. This information will provide an unprecedented molecular mechanistic insight and identify novel therapeutic targets for RTT that are likely relevant to multiple ASDs. 

(Year 1 grant)

Cortisol is the main stress hormone in human beings. Recently, autism research has gained interest in cortisol as a biomarker to characterize the activity of a significant component of the neuroendocrine system in autistic patients and fetuses who eventually grown into autistic patients. This project seeks to develop a synthetic receptor for cortisol that will serve as the basis for a continuous, real time sensor for cortisol. Such a sensor would collapse the human cost for cortisol measurements, as well as enable data collection that would otherwise be impractical. Such data collection would elucidate some underlying mechanistic features of autism and may revolutionize diagnosis and treatment. 

(Year 1 grant)

Effective social interactions are essential for human well-being. Individuals with Autism Spectrum Disorders experience disproportionate difficulties with social interactions, causing isolation for patients, suffering for families, and extensive societal costs. However, in spite of decades of effort, we do not yet know what aspects of brain functioning explain these disproportionate deficits. We propose a novel hypothesis: that deficits in social interaction reflect reduced neural flexibility of social information processing. We have designed a task that separately measures neural responses to dynamic emotional facial expressions that are driven by the external stimulus (e.g. the appearance of the face) versus those that are determined by the observer’s internal goals (e.g. to pay attention to the person’s emotion, while ignoring her age). In a pilot study, we found that in control adults, patterns of neural responses in many regions of cortex reflected the participants’ internal goals, rather than the external stimulus. We hypothesize that in individuals with Autism, the influence of flexible internal goals will be reduced, while the influence of the external stimulus will be preserved or even enhanced. Reduced neural flexibility of social processing may help to explain the impairments experienced by individuals with ASD, in real world social interactions.

(Year 1 grant)

In this exploratory project we aim to mathematically model the limits/capabilities of the different learning algorithms used by autistic vs. typical brains, when it comes to perceiving and processing inputs characterized by qualitatively different complexity, such as human emotions. Our goal is i) to interpret known to-date limitations of autistic brains in decoding inputs with specific qualitative characteristics (e.g. facial expressions and voice alterations related to basic and complex emotions), and ii) to design experimental procedures and measures that will rigorously quantify learning styles and assess/predict their capability to perceive such inputs. A longer-term goal involves the design of optimal training procedures that will take into account the characteristics of the individual’s learning style.

(Year 1 grant)

Uncontrolled Inflammation during pregnancy has been proposed as one environmental risk factor to the development of autism spectrum disorder (ASD) in affected children. Using a well-established rodent model mimicking this pathophysiology, we have recently identified a particular type of immune cells and their secreted proteins as direct mediators of inflammation-induced ASD-like phenotypes. We have also found disorganized cortical development in the affected fetal brain. In this proposed study, we aim to gain more insights into how inflammation affects brain development by identifying differentially regulated genetic factors and by developing novel genetically engineered mouse lines.

(Year 1 grant)

Infants spend most of their time looking at, and interacting with, other people. In natural scenes, typically developing infants’ attention is strongly drawn to familiar or friendly people, and to human faces, eyes and hands. By contrast, infants at high risk for Autism Spectrum Disorders (ASD) show disruptions in these very early signatures of social attention. Differences in early social attention may be critical to the developmental course of ASD, because where infants direct their attention determines both their social interactions and their opportunities for learning. However, the mechanisms driving infants’ social attention are poorly understood. Functional near-infrared spectroscopy (fNIRS) is a non-invasive neuroimaging method that can measure blood flow related to neural activity in awake infants sitting on their parents’ laps. We propose to use fNIRS to test two possible drivers of social attention: do infants pay more attention to people (or parts of people, such as eyes) because they “like” them (i.e. find them rewarding) or because they are more “interesting” (i.e. a better source of novel information)? Do each of these mechanisms drive social attention under different circumstances or for different aspects of the social world?

(Year 2 grant)

Autistic brains are distinguished in part by their “functional connectivity” long distance correlations in the neural activity of different brain regions. The meaning of functional connectivity patterns and how they relate to the behavioral phenomena of autism are still unclear, and investigation of functional connectivity mechanisms in humans is not possible. This proposal will address this problem by examining functional connectivity in a rat model of fragile X syndrome, a genetically defined autism spectrum disorder. Brain imaging experiments in animals will be used to characterize the differences between normal and diseased rats. Localized inactivation experiments will then be performed to test whether specific brain regions can account for functional connectivity effects, and to study how connectivity networks depend on the contributions of individual neural structures. This work will advance the study of autism by probing the pathology of an animal model and by suggesting specific neural activity manipulations that could either simulate or potentially reverse the physiological changes that occur in autism disorders.

(Year 1 grant)

Individuals with autism are known to have difficulties connecting with other people, reciprocating social interactions, and being emotionally regulated by others. Yet, until recently, very little attention has been given to the way people interact together, in a system, rather than by themselves. We propose a new way to collect data on how caregivers and their children, with and without autism, affect and are affected by each other (i.e., how they ‘sync-up’ with one another), both in their behavior as well as in their physiology. We also introduce a customizable digital-physical smart toy platform that will allow us to test hypotheses and collect data about patterns of caregiver-child synchrony in a naturalistic and engaging environment. Together, we hope to forge a new collaboration between smart toy technology and autism research that will help uncover how the social brain develops normally and might be differentiated in autism.

(Year 2 grant)

A growing body of work suggests that multi-level attentional factors may contribute to both the social and the non-social symptoms of ASD; fundamental differences in the weighting and prioritization of environmental information, especially during development, influences the way individuals with ASD process visual scenes, learn language, approach novel surroundings, and navigate social situations. In these proposed studies, we will test basic attentional mechanisms in 1.5-3-year-old toddlers with and without ASD (the earliest age when the condition can be reliably diagnosed) using both behavioral (gaze direction) and physiological (pupil dilation) measures. We propose that differences in early attentional engagement at the task- (but not stimulus-) level are a core factor of ASD etiology.

(Year 1 grant)

Difficulties in interacting with dynamic objects can have grave consequences for autistic individuals. This project aims to investigate these abilities across a range of motor tasks to test whether the difficulties seen clinically result from an underlying impairment in temporal prediction. The project uses computer vision tools to create novel experimental protocols that yield rich quantitative data regarding movement of individuals relative to the objects in the environment, besides being safe and suitable for a broad spectrum of children. Our experimental suite spans motor tasks at different scales, ranging from fine eye-movements to gross whole-body motion. Deeper knowledge of the issue will have relevance for making environments safer for autistic children and adults, as well as for designing interventions that acknowledge and address the underlying neurocognitive issue (e.g., prediction), and not merely the manifestation of the underlying impairment (e.g., difficulty in catching a ball).

(Year 1 grant)

Individuals with autism report exceptionally quick and accurate perception of small details in the visual environment. Yet, this detailed perception is not always to an individual’s advantage: individuals with autism also report difficulty in “filtering out” redundant and conflicting sensory information, leading to a sensory experience which is often described as “overloaded”.

We have recently discovered a behavioral correlate of this hallmark autistic symptom: individuals with autism evidence strikingly weak perceptual inhibition of conflicting information during binocular rivalry (Robertson et al., 2013), a simple perceptual task which can be used to test the strength of inhibition between populations of neurons with opposite selectivity in visual cortex.

We now seek to establish the neurobiological origins of this effect using a combination of highresolution magnetic resonance spectroscopy (MRS), functional imaging methods (fMRI and EEG), and behavioral techniques. Critically, we aim to link these fundamental perturbations in autistic visual behavior to a prominent hypothesis of autistic neural circuitry: that concentration of a particular neurotransmitter, GABA, may be reduced in the autistic cortex.

Converging lines of evidence support the view that autism is a disorder of brain connectivity in which abnormal brain structure and reduced coordination of activity across brain regions give rise to core features. The findings of functional connectivity MRI (fcMRI) and diffusion tensor imaging (DTI) studies of the brain are major sources of evidence for this theory. Although these techniques can illuminate brain connectivity in vivo, they are highly sensitive to head motion. Recent studies demonstrate that head motion can give rise to the patterns of abnormal connectivity that have been reported in children with autism. Thus, prior findings of brain connectivity abnormalities in autism may instead reflect motion artifact, and future studies must overcome this problem to be valid reflections of brain connectivity in autism. We propose to develop functional and diffusion MRI as a valid research tools for children with autism, without requiring sedation or anesthesia, by incorporating millisecond level motion detection and correction into faster pulse sequences, and developing an optimized data analysis pipeline. This will ameliorate a major source of artifact and speed the scanning process, thereby maximizing the chance of successful scans. These tools will benefit current and future studies of children with autism.

Autistic brains are distinguished in part by their “functional connectivity” long distance correlations in the neural activity of different brain regions. The meaning of functional connectivity patterns and how they relate to the behavioral phenomena of autism are still unclear, and investigation of functional connectivity mechanisms in humans is not possible. This proposal will address this problem by examining functional connectivity in a rat model of fragile X syndrome, a genetically defined autism spectrum disorder. Brain imaging experiments in animals will be used to characterize the differences between normal and diseased rats. Localized inactivation experiments will then be performed to test whether specific brain regions can account for functional connectivity effects, and to study how connectivity networks depend on the contributions of individual neural structures. This work will advance the study of autism by probing the pathology of an animal model and by suggesting specific neural activity manipulations that could either simulate or potentially reverse the physiological changes that occur in autism disorders.

(Year 2 grant)

Phelan-McDermid syndrome (PMS) is a genetic disorder in which a large proportion of patients present with autism symptoms. Since one of the genes missing in PMS, termed SHANK3, has also been found mutated in individuals with autism in the general population there is great interest in the study of autism in PMS. However, previous studies in PMS have not looked at genes close to SHANK3 and many cases have questionable diagnoses of autism. We intend to address these uncertainties by studying SHANK3 and neighboring genes in patients with PMS with well-established diagnosis of autism. 

(Year 2 grant)

The majority of ASD is not associated with a single genetic variant. Genome-wide association of multiple rare variants is plagued by low signal to noise in genetically heterogeneous backgrounds. In principle, functional genomics allows testing of putative rare variants against a uniform genetic background in vitro. However, while activating and inactivating single genes up to genome-scale is feasible with existing technology, efficient screening of multiple interacting genes for important neural phenotypes is not. Here we propose to establish proof-of-concept for a new approach that screens a diverse set of multi-genic perturbations for complex phenotypes by microscopy, and allows optical identification of the genotypes from the same image dataset.

Development of therapies to improve neurocognition in Autism Spectrum Disorders (ASD) requires understanding of this disease at cellular level. We have shown that in ASD patients with tuberous sclerosis complex, brain tissue is characterized by hypomyelination of axons, enlarged neurons, and disorganized neuronal connectivity. These cellular abnormalities likely contribute to cognitive deficits in ASD. We have started unraveling the possible biochemical causes behind these structural changes and identified potential drugs to improve neurocognition in ASD. However, the changes and effect of mechanical factors on neuronal connectivity and cognition in ASD have never been investigated. Based on our findings that hypomyelination causes decrease in tissue stiffness and demonstrated in vitro dependence of neuronal connectivity and branching on material stiffness, we hypothesize that the mechanical properties of the autistic brain may differ significantly from those of non-autistic brain, causing changes in structural and functional connectivity of neurons. To test this hypothesis, we propose to quantify for the first time the cell-scale mechanical properties of autistic brain from rodent models to determine the effect of mechanical environment on neuronal connectivity. Our findings could open new strategies for autism therapies, targeting mechanical cues, and provide novel diagnostic approaches based on stiffness detection with techniques such as magnetic resonance elastography.

Autism spectrum disorders (ASD) are typically identified by impaired language/communication, lack of sociability and repetitive behaviors. Here we present a new hypothesis that could explain these and other more recently identified symptoms of ASD. Specifically we propose that ASD symptoms arise from a failure of brain circuits to properly reinforce normal behaviors. Early social and language development depend critically on an infant orienting to social and verbal cues. Improper reinforcement of the response to these early cues could trigger a cascade of actions that ultimately leads to a lack of sociability and repetitive movements. The basal ganglia (BG) and their connection to cortex are known to be involved in the reinforcement of learned behaviors and their malfunction has been implicated in ASD. However, how the BG functions in normal reinforcement is not yet understood. We propose a new model in which the BG receives a copy of motor commands generated elsewhere, and we hypothesize that this copy is fundamental to the BG learning mechanism. We will test this model with optogenetics, in-vivo electrophysiology and behavior in mice. If validated, this radical new model could transform our current understanding of BG function and help understand the role of action reinforcement in ASD.

Converging evidence from many laboratories suggests that the projections from the hypothalamus that release the peptide oxytocin in many regions throughout the brain play a central role in social perception, cognition, and behavior. Despite this, the basic anatomical organization of this circuitry, as well as the roles of its components in neural mechanisms of social interaction, are still mostly unexplored. We will use cutting-edge genetic techniques, both those that are already established and some that are actively being developed in our laboratories, to determine key structural, functional, and genetic properties of this critical brain system. These studies will provide important insights into the organization of the social brain and may suggest intervention targets for treating its disorders.

(Year 1 grant)

Phelan-McDermid syndrome (PMS) is a genetic disorder in which a large proportion of patients present with autism symptoms. Since one of the genes missing in PMS, termed SHANK3, has also been found mutated in individuals with autism in the general population there is great interest in the study of autism in PMS. However, previous studies in PMS have not looked at genes close to SHANK3 and many cases have questionable diagnoses of autism. We intend to address these uncertainties by studying SHANK3 and neighboring genes in patients with PMS with well-established diagnosis of autism. 

Recent genome-wide studies have implicated the microRNA miR137 and its target genes in both autism spectrum disorders (ASD) and schizophrenia. Single-nucleotide polymorphisms (SNPs) associated with the risk of ASD and schizophrenia risk-associated may increase the expression of miR-137 in the hippocampus. Our preliminary data has confirmed a number of pre-synaptic proteins to be direct targets for miR-137, and suggest that the overexpression of miR-137 in the hippocampus leads to distinct abnormalities of the presynaptic compartment, which manifest in impaired synaptic plasticity and learning and memory. The current application proposes to characterize the role of miR-137 in learning and memory and synaptic function, and to probe the mechanisms by which alterations in miR-137 activity lead to cognitive dysfunction.

(Year 1 grant)

A growing body of work suggests that multi-level attentional factors may contribute to both the social and the non-social symptoms of ASD; fundamental differences in the weighting and prioritization of environmental information, especially during development, influences the way individuals with ASD process visual scenes, learn language, approach novel surroundings, and navigate social situations. In these proposed studies, we will test basic attentional mechanisms in 1.5-3-year-old toddlers with and without ASD (the earliest age when the condition can be reliably diagnosed) using both behavioral (gaze direction) and physiological (pupil dilation) measures. We propose that differences in early attentional engagement at the task- (but not stimulus-) level are a core factor of ASD etiology.

A major part of animals’ sensory and cognitive abilities is used to recognize and behave appropriately with respect to socially relevant information, and the dysfunction of the “social brain” is associated with psychiatric illnesses such as autism spectrum disorders. A number of studies have established oxytocin, a peptide released by neurons in the hypothalamus, as playing a key role in multiple aspects of social cognition, but the circuitry and mechanisms of its projections and pathways remain poorly understood. Here we propose to determine the fundamental structural and functional organizing principles of this central regulatory system of the social brain.

It is estimated that nearly 90% of all children on the autism spectrum suffer from sensory abnormalities, often hypersensitivities, to stimuli that a neurotypical individual could easily ignore (Leekam et al., 2007). These hypersensitivities can, in principle, be caused by abnormally acute sensory capabilities. However, empirical data contradict this possibility; individuals with autism do not differ systematically from neurotypical controls in their sensory acuity (DePape et al., 2012; Bölte et al., 2012). Here we consider an alternative account of hypersensitivities in autism. We hypothesize that autism is associated with reduced habituation, which in turn leads to reduced stimulus suppression. Immersion in an unrelentingly salient stimulus would lead to a sense of being overwhelmed. In this proposal, we describe experiments to rigorously test this hypothesis across multiple sensory modalities, and present preliminary data. This hypothesis can have significant impact on our understanding of a prominent aspect of the autism phenotype, and also help design novel kinds of early diagnostic tests.

Autism is a disease characterized by many inherited genetic mutations. Some of these mutations modulate brain function through direct alteration of codes for proteins in disease related biochemical pathways; one example is called an “Insulin-like Factor-1.” Mutations can also act indirectly through biochemical processes that modify a protein structure so that its activity is inactivated. One such type of modification is to attach a chemical group called nitric oxide to a sulfur group in a protein.

Nitric oxide is likely to be involved through silencing of PTEN which controls pAkt synthesis in the IGF-1 pathway which has been shown to be critical in Rett Syndrome. Our lab has now accomplished a method for detecting and analyzing proteins and small molecules with this modification. The modes of analysis include both imaging and a powerful tool called mass spectrometry, so we can locate where these changes take place in a neuron and at the same time identify the specific molecules that have been transformed. In preliminary results we have already shown that in a normal mouse different regions of the brain have different amounts of this modification. Our goal for this proposal is to determine whether there are differences between a normal mouse and an autistic mouse, and then to identify the proteins that make that difference. This approach is synergistic with those that employ genetics and can identify new therapeutic targets for drugs to combat autism.

One of the key challenges in understanding brain function and misfunction is the extraordinary diversity of cells in the brain. Approaches that measure for example the level of every type of RNA have until recently been applied to “homogenized” samples – in which the contents of all the cells are mixed together. This has greatly limited our ability to use such techniques to understand human brain function and pathology. In the past year, new technologies have begun emerging to conduct such measurements in single cells, but they are not yet scalable to large numbers of cells, and are very costly. Here, we will develop a method to profile the RNA content of thousands of individual brain cells, quickly and inexpensively. To do so, we use special devices to encapsulate each cell in an individual drop, tag the RNA of each cell with a unique ‘molecular barcode’, measure the expression level of each RNA with sequencing, and then use the molecular barcodes to determine which cell each RNA molecule came from. We will use this approach to better understand both a mouse model of autism and samples from human brain.

Converging lines of evidence support the view that autism is a disorder of brain connectivity in which abnormal brain structure and reduced coordination of activity across brain regions give rise to core features. The findings of functional connectivity MRI (fcMRI) and diffusion tensor imaging (DTI) studies of the brain are major sources of evidence for this theory. Although these techniques can illuminate brain connectivity in vivo, they are highly sensitive to head motion. Recent studies demonstrate that head motion can give rise to the patterns of abnormal connectivity that have been reported in children with autism. Thus, prior findings of brain connectivity abnormalities in autism may instead reflect motion artifact, and future studies must overcome this problem to be valid reflections of brain connectivity in autism. We propose to develop functional and diffusion MRI as a valid research tools for children with autism, without requiring sedation or anesthesia, by incorporating millisecond level motion detection and correction into faster pulse sequences, and developing an optimized data analysis pipeline. This will ameliorate a major source of artifact and speed the scanning process, thereby maximizing the chance of successful scans. These tools will benefit current and future studies of children with autism.

A critical question in autism research is whether different genetic causes converge on a defined number of pathophysiological mechanisms. An approach pioneered by the SCSB is to use a common battery of tests to compare mouse models of autism. Of particular interest is the question of whether alterations in synaptic connectivity, function, and plasticity are at the core of autism pathophysiology. The tool best suited to address this question is the whole-cell patch clamp technique with which multiple aspects of excitatory and inhibitory synaptic currents, cellular excitability, and interneuronal connectivity can be characterized. However, this method has historically been shown to have extremely low throughput because of the high level of expertise required of the experimenter to achieve good recordings from connected neurons. We propose to overcome this roadblock by developing a new image-guided 3D automated patchclamp system (Autopatcher 3D) for brain slices, extending the “blind” in vivo automated wholecell patch clamp prototype previously developed in the Boyden laboratory (MIT) collaborating with the Forest Laboratory (Georgia Tech). This system will provide high throughput, multi-cell, patch clamp electrophysiology for analyzing functional neuronal connectivity in vitro, as well as allowing for the measurement of neuronal responses in vivo. We will use Autopatcher 3D to test whether there are any changes in the functional connectivity of the visual cortical microcircuit, and how they may lead to enhanced persistent activity in Fmr1 KO mice.

Learning where and how neural changes associated with autism take place is critically important for understanding the physiological basis for autism spectrum disorders and relating genetic abnormalities to brain function. Noninvasive brain imaging methods offer a promising route to accomplishing this. In this proposal, we describe efforts to perform noninvasive MRI of calcium ions, which are thought to be particularly closely related to autism pathologies. Our proposal includes the synthesis of innovative calcium-sensitive MRI contrast agents, as well as the application of these new sensors and related imaging techniques to studies of autism-related pathology in a mouse model. The new imaging methods we are developing could be applied in additional animal models, and ultimately perhaps for clinical investigation and diagnosis in humans.

Our goal is to develop new machine-learning-based techniques for behavioral phenotyping of mice, with direct application to autism. By automating the phenotyping process, such techniques offer the benefits of speed, objectivity and standardization across labs. More specifically, our proposed work has two components:
1. Automated phenotyping of social mouse behavior. Social abnormalities are amongst the core phenotypes characterizing mouse models of autism [4,5]. We will develop a fully automated computer system for phenotyping social mouse behavior, to be applied to autism studies.
2. Multi-dimensional behavioral phenotyping. Mouse models of autism are characterized along a number of behavioral dimensions, necessitating multiple behavioral assays [4,5]. We believe that quantitative measurements derived from machine learning might be sensitive to many of the behavioral variations currently tested and might even provide a richer characterization of behavior. We will investigate the feasibility of a simplified alternative to at least some of the existing behavioral assays, such as light/dark exploration and the Morris water maze, leveraging machine learning to capture and quantify mouse behavior along multiple dimensions.
This work builds on technology previously developed in the Poggio Lab, including a trainable system for phenotyping single mice in their home cages [1] and preliminary methods for tracking multiple socially-housed mice [2,3].
References:
1. H. Jhuang, E. Garrote, X. Yu, V. Khilani, T. Poggio, A. Steele, and T. Serre. Automated home-cage behavioral phenotyping of mice. Nature Communications, 2010.
2. N. Edelman. Automated phenotyping of mouse social behavior. M.Eng. thesis, MIT, 2011.
3. S. Braun. Tracking multiple mice. M.Eng. thesis, MIT, 2012.
4. J. Peca. C. Feliciano. J. Ting. W. Wang. M. Wells. T. Venkatramann. C. Lascola. Z. Fu. and G. Feng. Shank3 mutant mice display autistic-like behaaviors and striatal dysfunction. Nature, 2011.
J. N. Crawley. Mouse behavioral assays relevant to the symptoms of autism. Brain Pathology, 17:448-459, 2007.

Autism Spectrum Disorder (ASD) is a common pediatric disease with a prevalence of at least 1%. We propose here to provide a scalable, reproducible mechanism to ascertain and characterize populations with ASD using software running against a wide variety of electronic health record systems. Our aims are to create a set of gold standard corpora for enabling comparison of ASD patients with controls matched on age, gender and non-ASD related morbidity. We propose to build a predictor function using machine learning techniques on human-curated gold standard data to identify patients with ASD and their clinical ASD subtypes. The function’s input data will come from both codified medical record data and natural language interpretation of the narrative records that are prevalent in clinical data. We will measure the performance of the predictor against our gold standard data and then against 14,000 ASD patient records in the Children’s Hospital i2b2 data repository and over two million other patients from among which we will select controls. This process will yield a much larger automatically identified cohort of ASD patients and controls, to support further research on ASD. We also anticipate that other institutions will apply our methods to generate additional cohorts that can be used to validate findings or to create larger merged research data sets that are consistently identified.

We identified a C. elegans gene that encodes a novel regulator of synaptic release, C44B9.1, and demonstrated that its mammalian counterpart, CCDC132, is conserved structurally and functionally. C. elegans C44B9.1 mutants have a behavioral and pharmacological profile identical to that of other worm mutants defective in the C. elegans orthologs of mammalian genes involved in neurotransmission and associated with autism spectrum disorder (ASD). The human CCDC132 gene is located in a tiny chromosomal region associated with ASD. These observations indicate that CCDC132 is a plausible candidate ASD gene. This project will determine the levels of expression of murine CCDC132 mRNA and protein in different regions of the brain, at different developmental stages and in different neuron types. Proteins that interact with CCDC132 will be sought, and the effects of CCDC132 knockdown on synaptic and dense-core vesicle function will be determined. The effects of C. elegans C44B9.1 on postsynaptic proteins and processes also will be assessed. A mouse CCDC132 knockout mutant will be constructed. 

We propose to create the first comprehensive snapshot of the molecular pathways affected in Angelman syndrome (AS) using human neurons derived from UBE3A knock-out embryonic stem cells. This project will bring together cutting-edge technologies from two labs: isogenic disease modeling using novel genome engineering technology coupled with rapid, efficient differentiation of human neuron subtypes (Zhang lab) and high-resolution quantitative proteomics profiling of UBE3A (Carr lab). Our goal is to uncover the molecular basis of synaptic dysfunction in AS and identify possible targets for therapeutic intervention.

Defining molecular pathways that are dysfunctional in autism and autistic spectrum disorders (ASDs) is key to understanding their pathogenesis and developing future therapeutics. In Alzheimer’s and Parkinson’s Disease, identification of single gene mutations in the 5-10% of genetic cases have revealed core molecular pathways that are altered, including in the larger category of sporadic cases. A key question is whether a similar molecular pathway(s) will emerge for autism based on the recent identification of defined mutations and de novo genome copy number variations that account for 10-20% of ASDs. Here, we propose to take advantage of genetic manipulations available in Drosophila to explore the mechanisms by which the autism-associated endosomal protein, NHE9 (Na+/H+ exchanger 9), couples alterations in neuronal activity to modifications of synaptic connectivity. NHE9 is one of several newly identified genetic links that indicate abnormal endosomal trafficking and synaptic growth may predispose to autism. Work from my lab has recently characterized a synaptic growth and plasticity pathway in Drosophila where postsynaptic targets release retrograde signals in an activity-dependent manner, triggering synaptic maturation and growth. These growth signals are processed and regulated through trafficking in the presynaptic endosomal pathway. We hypothesize that mutations in NHE9 alter endosomal formation, or ligand-receptor association in endosomal compartments, through disruption of endosomal pH, leading to abnormal synaptic growth signaling and activity-dependent defects in brain wiring that might contribute to autistic behavior. Similarly, postsynaptic glutamate receptors cycle in and out of the postsynaptic membrane through an endosomal pathway, resulting in acute changes in synaptic strength that may also require NHE function. Endosomes exhibit a progressive acidification from early endosomes (pH ~6.5) to lysosomes (pH ~4.5) that is essential for degradation and recycling of internalized ligand-receptor complexes, transmitter receptors and cell adhesion proteins. Together with the vacuolar V-ATPase, NHE9 is predicted to be the key molecular determinant of endosomal pH by allowing early signaling endosomes to remain relatively basic by transporting H+ ions out of this compartment, where ligand-receptor pairs can remain attached and transmit synaptic growth signals. We will determine if endosomal trafficking defects are present in NHE9 mutants, and whether NHE9 activity may be linked to other ASD mutants that include synaptic cell adhesion proteins like Neurexin and Neuroligin. The generation of new contact sites is likely to require removal or recycling of surface Neurexin and Neuroligin through the endosomal system, a process that may require NHE9. Linking a common synaptic defect in these seemingly unrelated proteins may reveal a conserved molecular pathway that is dysfunctional inautism. Defining the underlying synaptic pathology of autism-linked loci using a model system like Drosophila fits the overall mission of the Simons Center to link molecular causes of autism with the underlying pathophysiology of this neurodevelopmental disorder. 

The hypothesis we propose to develop and test in this project is that some salient aspects of the autism phenotype, such as insistence on sameness, may be manifestations of an underlying impairment in predictive abilities. Our goal is to design and conduct behavioral and electro-physiological (MEG and EEG) tests to assess performance of typically developing (TD) and ASD participants on tasks with systematically varied levels of predictive difficulty. The results have the potential of providing unifying mechanistic insights into multiple aspects of autism.

It is well known that many people with autistic spectrum disorders prefer communication via online social networks to face-to-face verbal communication [Blume 97]. However, to date, little attention has been paid specifically to designing social network software that meets their needs. This proposal aims to use artificial intelligence technology, machine learning, common sense reasoning, and affective sensing to make communication via social network software more effective and more satisfying for people on the autistic spectrum.

The goal of the proposal is to use functional magnetic resonance imaging (fMRI) to address a pressing and medically important issue in the treatment of patients with autism spectrum disorders (ASD). The most common comorbidity in highfunctioning ASD is ADHD (in as many as 67% of patients), but there is great uncertainty for physicians, patients, and families as to whether the inattention or hyperactivity shown by so many ASD patients (and which earns them a comorbid diagnosis of ADHD) is the same neurobiological disorder as ADHD, or is, instead, a different disorder altogether that appears like ADHD, but is really a more direct expression of ASD that only superficially mimics ADHD. This uncertainty, which is hotly debated among physicians who treat ASD patients, is important because the answer guides what treatment is given to many ASD patients (i.e., whether or not the patients are chronically prescribed methylphenidate or other medications). The proposed study would provide the first scientific evidence as to whether ADHD in the context of ASD is simply the combination of two disorders (and should be treated like ADHD) or is instead a fundamentally different disorder (and should not be treated like ADHD).

Genetic factors behind autism remain poorly understood. We propose an integrative analysis, covering both novel as well as public datasets, to identify the pathways and genes involved in autism. The Kohane lab has collected new gene expression datasets from autism patients as well as controls. The lab also has access to recently generated data on exomic variations in autism patients and controls. The Berger and Kohane labs propose to collaborate on an integrative analysis of this data, also including publicly-available data from SFARI. Our first aim is to construct multiple molecular subnetworks by mapping each autism-related dataset (expression, exomic variation, CNVs) to protein-protein interaction (PPI) networks. The Berger lab has extensive experience in the statistical and graph-theoretic techniques required for constructing robust networks from such data. Our second aim is to compare, combine and contrast these subnetworks. By bringing our expertise in network alignment to the problem, we aim to discover the consensus network of genes and pathways involved in autism that best agrees with the combined data. Furthermore, we will repeat this analysis for various subclasses of the disease (PDD, Asperger’s, etc.) to identify genetic factors common across multiple subclasses as well factors specific to each.

The primate basal-ganglia circuit originating from the pregenual anterior cingulate cortex (pACC) has been considered to play an important role in outcome evaluation as well as control of emotion and motivation. Our laboratory previously found that the primate pACC preferentially projects to the striosome compartment of the striatum, and that neuronal activity in the pACC biases decision-making by producing an anxiety-like state. Interestingly, this medial prefrontal region overlaps the region that has classically been implicated in vocalization in non-human primates, suggesting a potential role of the pACC in emotional control in socially conflicting states as well as social attachment. We propose to focus on the possibility that the basal-ganglia circuit originating in the anterior cingulate and medial prefrontal cortex may help to control avoidance and approach in social behaviors whose more immediate neural basis lies in regions such as the hypothalamus and amygdala.
To address this hypothesis, we will first perform a pilot study to identify specific cortico-basal ganglia circuitry that controls social attachment in monkeys. We will test this hypothesis by examining the changes in pair-housed monkeys’ social interactions in freely-moving conditions while we apply microstimulation to the pACC using a wireless device. To monitor the possible behavioral changes induced by this brain stimulation, we will simultaneously use a wireless telemetry sensor system that enables us to record autonomic and emotional responses such as skin conductance and heart rate. Second, in a laboratory experimental setup, we will determine whether striatal neurons, targets of the pACC, are responsive to positive or negative emotion, and we will attempt to identify the connectivity of this pACC-basal ganglia circuit using antidromic microstimulation and collision tests. The results of these two experimental approaches will be integrated to characterize the function of primate limbic cortico-basal ganglia circuitry in social valuation.

Loss-of-function mutations in the MECP2 gene account for vast majority of cases of Rett Syndrome (RTT), an X-linked neurodevelopmental disorder that belongs to the autism spectrum disorders. Most patients are females, carrying a single copy of mutated MeCP2 gene. Random X-chromosome inactivation (XCI) leads to cellular mosaicism in which cells express either the wild-type or mutant MeCP2 allele. Recent landmark studies in RTT mouse models demonstrated that re-expressing MeCP2 gene in symptomatic animals could reverse disease progression and restore functions. These findings strongly suggest that RTT is a treatable disease, and raise the intriguing possibility that reactivating the wild-type MeCP2 gene on the inactivated X-chromosome could confer therapeutic benefits. It has been reported that certain carriers of MeCP2 mutation are only mildly affected or in fact symptom-free due to favorable skewing of XCI, which leads to expression of the wild-type MECP2 allele and silencing of the X-chromosome on which the mutant copy resides. This observation supports our hypothesis that increasing the ratio of cells that express the wild-type MeCP2 gene can ameliorate RTT symptoms. To identify genes and drugs and possible drug targets that activate the silent wild-type MeCP2 gene, we propose to develop and implement high-throughput chemical genetic and genome wide functional screening by means of isogenic human cells expressing fluorescent markers from the endogenous MeCP2 loci. Our proposed project will establish a novel platform for therapeutic discovery for RTT and possibly other X-linked neurodevelopmental disorders. To our best knowledge, this will be the first targeted drug and gene screen for RTT using isogenic human cells.

The aim of this Simons Seed Grant project is twofold: A) to develop a precise formal model of how we might reason about agent’s goal-directed actions, and B) to test the implications of this model behaviorally in both typically developing children and children with ASD. Motivated by the formal rational model, we will be in a better position to pinpoint specific aspects of social reasoning that are preserved or compromised in ASD. This in turn can constrain our search for the biological mechanisms underlying this disorder.
Specific Aims are as follows:
1. Develop a computational model of social evaluation to better understand the roles of competence, constraints and motivation in inferences about agents’ goal-directed actions.
2. Investigate typically developing children’s ability to draw the inferences predicted by the model, and whether and how these inference capacities change over development.
3. Compare the inferences of children diagnosed with ASD with both the model predictions and the results of typically developing children in order to identify the precise nature of impairments in social cognition in high-IQ children with ASD.

Restricted repetitive behaviors (RRBs) are an important symptom of autism and other CNS disorders. Research on both humans and animals has implicated abnormalities in the cortico-striato-thalamo-cortical (CSTC) circuit. According to the direct-indirect imbalance hypothesis, RRBs result when the direct pathway of the CSTC circuit overpowers the indirect pathway. According to the extreme habit hypothesis, RRBs result when excessive activity-dependent plasticity causes learning of extreme habits. Studies of the Shank3 mouse model of autism point to an abnormality of corticostriatal connectivity, suggesting that these two hypotheses could be tested by a closer examination of cortical connections onto medium spiny neurons (MSNs) of the striatum. We propose to do this by using serial electron microscopy (EM) to reconstruct MSNs and their synapses from cortical afferents. Serial EM will be preceded by light microscopy (LM) to distinguish between direct versus indirect pathway and active versus inactive MSNs, using fluorescent labeling with fos-GFP and D1R-tdTomato. Shank3 mutant mice will be compared with wild type mice. The specific aims for this seed grant are to demonstrate the feasibility of the study by acquiring LM and serial EM images of striatal volumes, registering them with each other, reconstructing MSNs and their cortical afferents, and quantifying connectivity.

Our research study seeks to understand how the flow of knowledge across the social networks of health care providers can influence the screening, diagnosis and reported prevalence of Autism Spectrum Disorder (ASD). We hypothesize that physicians facing the complex challenges of diagnosing autism may be influenced by the knowledge and opinions of other respected physicians. Over time, this process can result in the propagation of diagnostic knowledge and practices predisposed towards ASD throughout the population of physicians.
We will test our hypothesis using data from Kaiser Permanente North California, the California Department of Developmental Services, and California birth certificate data. Based on the electronic medical records and personnel files at Kaiser Permanente, we will construct measures of social learning between health care providers. This measure will be included along with measures of other more traditional risk factors (e.g. sex, maternal education) in a discrete-time survival model of the referral and diagnosis of ASD. By assessing the effect of social networks of health care providers, this study can help explain the increasing the reported prevalence and geographic clustering of ASD.