New research provides a potential brain-based explanation for social touch aversion in some forms of autism. A study published in Nature Communications finds that in a mouse model of Fragile X syndrome, a leading genetic cause of autism, neurons in key sensory and emotional brain regions fail to differentiate between being touched by another mouse and being touched by a plastic object. This apparent neural confusion is mirrored in the animals’ behavior, as they treat both social and non-social interactions as equally unpleasant, especially when the contact is unsolicited.
For most social animals, including humans, touch is a fundamental channel of communication and bonding. From a comforting hug to a friendly pat on the back, physical contact can convey emotion, build relationships, and offer solace. Brain circuits have evolved to recognize the unique significance of social touch, often prioritizing it over contact with inanimate things. However, for some individuals with neurodevelopmental conditions like autism spectrum disorder, social touch can be perceived not as pleasant, but as overwhelming or aversive.
This difference in sensory experience is a major focus for scientists seeking to understand the neural basis of autism. A team of UCLA researchers, led by Trishala Chari and Carlos Portera-Cailliau, set out to investigate the neural underpinnings of social touch aversion. They started with a specific question: Could it be that for individuals who find social touch unpleasant, their brains are not processing the unique “social value” of that touch? Perhaps, at a neural level, the brain fails to distinguish between a social interaction and a non-social one, leading to a negative or avoidant response to both.
“My lab is interested in understanding what is different at the circuit level in neurodevelopmental conditions. Fragile X syndrome is the most common inherited cause of intellectual disability and the leading single‑gene contributor to autism, and there is a very good mouse model for it (because it reproduces many of the same symptoms of the human disease), so we have focused on it,” explained Portera-Cailliau, a professor of neurology and neurobiology at the David Geffen School of Medicine at UCLA.
“I like to say we follow a symptom-to-circuit approach, meaning that we start with a specific symptom of autism, like sensory hypersensitivity and then record from neurons in circuits that would be responsible for that symptom. For example, in the past we discovered that aversion to repetitive whisker stimulation in Fragile X mice is associated with reduced habituation of neurons in the somatosensory cortex that process those whisker stimuli, whereas in control healthy mice (wild-type mice) the same neurons reduce their firing gradually because the whisker stimulation is not bothersome to them – it is non-threatening and non-salient – so they ignore it.”
“In this study, we wanted to investigate the circuit basis of another related symptom in Fragile X syndrome, aversion to social touch (hugging, kissing, tickling),” Portera-Cailliau said. “The ultimate goal of our symptom-to-circuit approach to neurodevelopmental conditions is to eventually identify an intervention at the neuronal level that might restore circuit function to wild type control levels, as a therapy for that symptom.”
To explore this, the researchers used a special strain of mice, known as Fmr1 knockout mice, which exhibit behaviors that are similar to Fragile X symptoms in humans, including sensory hypersensitivity and social avoidance. The scientists aimed to connect a specific behavior—aversion to social touch—directly to the activity of neurons in the brain circuits responsible for processing it.
The research was conducted in two main phases. First, the researchers needed to identify which brain regions were most active during social touch. To do this, they used a sophisticated genetic tool in a group of wild-type mice. This technique permanently marks any neurons that become highly active during a specific time window.
The researchers placed these mice in an apparatus where they were held comfortably in place but could run on a ball. A motorized platform then repeatedly presented them with either another mouse for a social interaction or a plastic tube for a non-social one.
By analyzing the brains of these mice afterward, the scientists could create a map of the brain regions that became active in response to these different types of touch. This mapping confirmed the involvement of the somatosensory cortex, a region that processes touch from the whiskers, and also highlighted two other key areas: the basolateral amygdala, known for its role in processing emotions like fear and pleasure, and the tail of the striatum, which is involved in making decisions based on sensory information.
In the second, and primary, phase of the study, the scientists conducted a more complex experiment with both wild-type mice and the Fmr1 knockout mice, the model for Fragile X syndrome. They surgically implanted state-of-the-art Neuropixels probes into the brains of these mice. These incredibly fine probes contain hundreds of electrodes, allowing the researchers to simultaneously record the individual electrical signals, or “spikes,” from hundreds of neurons across all three target brain regions: the somatosensory cortex, the tail of the striatum, and the basolateral amygdala.
With these probes in place, the mice were again placed in the behavioral apparatus. This time, the experiment was designed to test two different factors. The first was the context of the touch: social (another mouse) versus non-social (a plastic object). The second was the element of choice. In “voluntary” interactions, the platform would bring the other mouse or object close enough for the test mouse to reach out and explore it with its whiskers. In “forced” interactions, the platform moved closer, bringing the other mouse or object into direct contact with the test mouse’s snout, invading its personal space.
Throughout these trials, high-resolution cameras recorded the mice’s faces and bodies. The researchers used specialized software to analyze these videos for signs of aversion, such as running to avoid the interaction, squinting the eyes (orbital tightening), or moving the whiskers in a specific way.
“Our novel assay also allowed us to compare situations in which the interactions were voluntary (meaning the test mouse could choose to interact by whisking to palpate the other mouse or object) and situations where the test animal was forced to interact because the other mouse or object was brought into direct contact with its snout (basically being brought into its ‘personal space’ which we imagined would be more aversive than voluntary interactions),” Portera-Cailliau explained.
“This assay is pretty sophisticated: it involves a motor that moves the platform where the visitor mouse or object is placed, several high resolution cameras that focus on the mouse’s head, pupil, body, a bunch of electronics to keep track of when things happen, and custom code written in MATLAB and Python to run it all smoothly. Trishala built it all from scratch.”
The findings revealed a stark contrast between the two groups of mice, both in their behavior and their underlying brain activity. The wild-type mice behaved as expected for a social animal. They showed clear signs of aversion to the inanimate object, especially when it was forced upon them. They would run to avoid it and display negative facial expressions. When presented with another mouse, however, they were much more tolerant and did not show these same aversive behaviors. Their actions indicated that they could distinguish between a meaningless object and a meaningful social partner.
The Fmr1 knockout mice behaved very differently. They failed to make this distinction. They showed just as much aversion to the social touch of another mouse as they did to the non-social touch of the plastic object. Their behavior suggested that both experiences were equally unpleasant to them. This was particularly evident during forced interactions. When social touch was forcefully imposed onto them within their personal space, the Fmr1 knockout mice showed more aversive facial expressions than the wild-type mice, indicating that unsolicited social contact was especially bothersome for them.
“Wild-type mice clearly differentiated between social and non-social stimuli because they showed aversion to being brought in contact with an object, but not a mouse,” Portera-Cailliau told PsyPost. “We know this because we video-recorded their faces and used machine-vision approaches to analyze their facial expressions. Control mice showed squinting of their eyes and unusual posturing of their whiskers (whisker protraction) only when presented with an object, but they tolerated social interactions very well. We interpret this to mean that control mice seem to enjoy social interactions, but don’t like repeated interactions with a plastic object or a furry toy mouse, particularly when we forced it onto their personal space.”
“In contrast, Fmr1 knockout mice showed similar degrees of aversion to both object and social touch (even voluntary touch), as if they couldn’t tell them apart or felt that social touch was just as unpleasant as non-social touch. In other words, we could reproduce what is seen in the clinic with human Fragile X syndrome patients in the mouse model.”
The recordings from the Neuropixels probes provided a direct window into the brain activity driving these behaviors. In the wild-type mice, the neurons clearly reflected the animals’ behavioral preferences. Neurons in the somatosensory cortex showed different patterns of activity for social versus object touch, with a general preference for social stimuli. This suggests the sensory cortex does not just register the physical sensation of touch, but also its context.
In the tail of the striatum and basolateral amygdala, activity seemed less about the identity of the stimulus and more about the choice; neurons in these regions fired most intensely during the most aversive condition for the wild-type mice: forced object touch. This suggests these areas are involved in processing unwanted or unpleasant interactions.
In the brains of the Fmr1 knockout mice, this organized and discriminating neural response was largely absent. Neurons in their somatosensory cortex and tail of the striatum were much less able to tell the difference between social and object touch, especially during voluntary interactions. The distinct neural signature that signified “social” in the wild-type mice was blunted or missing.
Using computational models, the researchers confirmed that they could accurately predict whether a wild-type mouse was experiencing social or object touch just by looking at its neural activity. For the Fmr1 knockout mice, this prediction was far less accurate, because the neural signals for the two conditions were so similar. This suggests that the behavioral inability to distinguish social value stems from an inability of the underlying brain circuits to represent that value.
Finally, the researchers used a linear encoding model, a computational method to determine how strongly the activity of a neuron is related to different behaviors. In wild-type mice, the activity of neurons in the somatosensory cortex was strongly linked not only to the type of touch but also to the aversive facial expressions the mice made. In the Fmr1 knockout mice, this relationship between neural activity and aversive behavior was weaker and altered. This finding reinforces the idea that in this model of autism, the entire chain of processing, from representing a social stimulus to generating a behavioral response, is different.
“We found that neurons in the somatosensory cortex of control mice fired differently to social touch compared to object touch,” Portera-Cailliau said. “Most neurons showed greater modulation of their activity with social touch than with object touch, meaning they fired more to social interactions.”
“In contrast, neurons from Fmr1 knockout mice fired more similarly to both presentations, implying that at the network level the brain could not discriminate between social and non-social stimuli. This was further confirmed with a linear encoding model that determined how well each behavior variable (eye squinting, whisker protraction, pupil size, etc.) was encoded by the activity of a particular neuron.”
“The take home message is that neurons in mice with Fragile X symptoms seem less able to differentiate between social and non-social interactions, which could explain why at the behavior level they find both similarly annoying/aversive. Therefore, social avoidance in Fragile X syndrome/autism could relate to differences in the activity of neurons. By tweaking that activity it may be possible to restore a preference of certain neurons for social touch, which could help individuals better tolerate social interactions they find uncomfortable.”
The study does have some limitations. While behaviors like facial squinting and avoidance are strong indicators of a negative experience, it is impossible to know for certain what a mouse is subjectively feeling. The results will also need to be reproduced in other laboratories and with other animal models of autism to understand how broadly they apply.
“One worry is that we sometimes fall into the habit of anthropomorphizing too much what we see in mouse behavior. We are interpreting the facial expressions of certain Fmr1 knockout mice to social touch as reflecting aversion, unpleasantness, discomfort, but we can’t tell for sure what they experience,” Portera-Cailliau noted.
“There is also a lot of variability between individual mice, with some showing a lot of aversion and others less, and there was no universal way by which mice showed aversion: some mice that showed a lot of squinting to social interactions showed little whisker protraction or running avoidance, and vice versa. This may not be surprising. Humans don’t always respond the same way to certain irritating stimuli (like nails on a chalkboard).”
“It will also be important for other labs to reproduce similar results, not just in Fmr1 knockout mice, but in other models of autism spectrum disorder,” Portera-Cailliau added.
Despite these caveats, the research provides a powerful insight into the brain mechanisms of social touch processing. The team’s next steps include recording from other brain regions that may be involved and testing whether existing drugs could restore a more typical pattern of neural activity. They also plan to enhance the experiment by giving mice more control, for instance by letting them press a lever to bring a social partner closer or keep them away. This could help explore more nuanced aspects of social preference.
The study, “A reduced ability to discriminate social from non-social touch at the circuit level may underlie social avoidance in autism,” was authored by Trishala Chari, Ariana Hernandez, João Couto, and Carlos Portera-Cailliau.
