Dissociated Identities (aka Alters) - How are they formed?

Is it possible to house multiple alternate selves within the same brain?  Dr. O'Leary argues that it is not only possible, but under certain conditions it is likely to happen.  Dissociative Identity Disorder is discussed not as a "shattering of the self,"  but instead as a failure to integrate the self during development.  Dr. O'Leary explores different biologically and computationally plausible models to explain how the self might fail to integrate during the normal process children go through in order to distinguish the self from others and others from others, as they build a (usually) coherent model of themselves.

Please leave feedback at https://www.psydactic.com or send any comments to feedback@psydactic.com.

References and readings (when available) are posted at the end of each episode transcript, located at psydactic.buzzsprout.com.  All opinions expressed in this podcast are exclusively those of the person speaking and should not be confused with the opinions of anyone else.  We reserve the right to be wrong.  Nothing in this podcast should be treated as individual medical advice.

Transcript

Copy Transcript Copied

SPEAKER_01 0:05

Welcome to Psydactic. Today is Friday, March 6th, 2026, and I am your host, Dr. O'Leary. I am a second-year child and adolescent psychiatry fellow in the National Capital Region. This is a podcast about psychiatry, psychology, and neuroscience. And today, I'm going to explore how the various altars in a person with dissociative identity disorder might be able to form during development. I produce this podcast alone, occasionally with the help of some AI. So everything I present here is my own responsibility, and nothing should be considered the opinion of anyone else, including the U.S. Federal Government, the Department of Defense, the Defense Health Agency, the Space Force, or my spouse. I take full responsibility for the content. A few years ago, someone asked me whether or not I believed in multiple personality disorder. The term multiple personality disorder is an historical and poorly descriptive name for what we now call dissociative identity disorder. At the time, I had not spent much time trying to understand it, and my response betrayed my ignorance. I don't see any plausible neurological mechanism that could explain it, so I'm not sure if it really exists. I was being honest. Since that time, I've changed my conception quite a bit. I am now confident that it is possible for the brain to operate in a way that could allow for alternate identities, with their own sets of defining memories and senses of self. It doesn't mean that each of these selves is equally deep or integrated, just that they are possible. Today I'm going to explore multiple ways in which our brain could create and house multiple identities. If you're looking for a definitive answer to everything about DID, I will disappoint you. But if you're looking to explore just how awesome and complex the process of reality construction and self-formation is in the brain, then buckle your seatbelt. But before I cut into the skull and expose the brain, I want to give a brief overview of how humans first started to suspect that other humans might house different identities within them. The oldest view I know of is that humans were thought to be possessed by various supernatural creatures, often called demons or spirits, and that these spirits took control of the body. Many afflictions of the body in the past have been attributed to supernatural beings, especially things that we couldn't see, the cause of, like fever or a delirium. One example of D.I.D. that can be picked from more modern history is that of Jean Ferry in 1584. She was a 25-year-old nun who recorded the details of her own exorcism. Her account, supplemented by the records of her exorcists, details the presence of multiple demons and devils, which we would now likely call altars. Each had its own name, identity, and specific functions. Jean's altars included what would probably now be classified as persecutory protectors, child altars, and internal self-helpers. Fairy described audible internal dialogue between these parts, periods where one part would take control of her body, and there was often amnesia for their actions. These actions ranged from rage to self-harm to disordered eating. Her condition, like many cases of DID, was linked to a history of childhood physical and possibly even sexual trauma and abuse. In fact, the most common cause of DID we know of is developmental trauma. Let me back up and talk about ways in which our brain can appear to house very different identities within them that are not actually fully dissociative. As a young man, I remember being dumbstrucked frequently when I would, for the first time, observe my friends around someone they had grown up with, especially a parent, their mother. My friend, who I had known to be thoughtful, empathetic, and calm, could turn into a demanding, mean-spirited, hateful, and obstinate child in the presence of their mother. Something about the context of the family made them act in ways that I had never seen them act outside of that context. What was more shocking to me was that they didn't even appear to notice that their behaviors were terrifying me. It was just perfectly normal to them. I also remember my father, whose own childhood was abusive, and who had then shipped off to Vietnam and experienced terrible things there. He later wrecked his car and suffered a severe traumatic brain injury that left him far more volatile and angrier than I had ever known him before. A small trigger could send him into a rage. Then he might just disappear for a few hours, occasionally for some days or weeks in the hospital, ashamed of his actions. I also remember Edu, who was one of the many homeless people that I spent many months with in New Mexico when I was a young man exploring my own identity. Eju grew up between various foster homes and claimed to have lived much of his childhood on an Indian reservation. He could take a small, flat rock, about the size of the padding of his thumb, and then flick it between his fingers so quickly that it would sing through the air. He claimed he could use it to hit a small animal with enough force to stun it, or even to kill it. I believed him because when that rock hit something, the sound was immense. Eju also claimed to have many names, and would be sure to introduce himself to you by whatever name he went by that day or hour. Probably all of us can think of examples, either of ourselves or of others, where there's been a dramatic shift in emotions and actions, resulting in someone saying or doing things that are completely out of character. I've had patients with premenstrual dysphoric symptoms so bad that they barely recognize themselves during those times. However, they don't lose time or memory for their actions. So it's very well established that most of us can feel and act at times like someone we normally aren't. Our brains are definitely capable of putting us into a state where we have little access to the kind of reasoning and remembered self that we mostly are. A manic person might suddenly decide that they're a gift to humanity. A depressed person may believe without any evidence that they are a burden to everyone. Neither of them will be able to call up a single memory that will argue with that point of view. But this is not the same as developmentally acquired dissociative identities. Everyone can at times relatively dissociate parts of themselves and have those parts kind of take over. When severe, like in the dissociative subtypes of PTSD, and in many cases of borderline personality disorder, which itself is often associated with early trauma, frequently there are more or less pronounced episodes that can be accompanied by uncharacteristic acts with partial or complete amnesia for what was happening. How did I end up on the floor in the janitor's closet? Oh shit. Why did I do that? What's wrong with me? Dissociative identity disorder is unique among our brains' abilities to take on and act in accordance with various identities because it results from a failure of the normal process of self-integration during early childhood. Somehow, this effectively walls off previous experiences from accessibility to the apparently normal self, who is generally the one interacting with the world. These walls can vary in how much they partition. Some of them might be leaky, resulting in what appear to be different voices or memories that don't belong to me. But these memories and voices aren't psychotic in origin. They're just relatively disparate integrated selves competing for conscious awareness. Like many psychiatric disorders, DID is so diverse and degenerate that I will not be able to do it complete justice today. What I want to accomplish is to demonstrate that DID is indeed a neurodevelopmentally plausible condition that in some respects inevitably follows from the way our brain constructs its sense of self and others. How the brain accomplishes this is unique in each individual, so please don't think that everything I describe here is necessarily universal among humans or humans living with dissociative identity disorder. Let's start at the very beginning, because I've been told that's a very good place to start. The problem is what is the beginning. I'm gonna define it today as birth because that's just convenient. It's a convenient way to begin to discuss what is the most important thing that a child can do before they develop a coherent sense of self. So to be a self, we must first define what is an other. Early psychoanalysts like Klein and Mahler developed something called object relations to help explain how a child individuates. They have internal self-objects and then develop external objects. One of the first tasks a child must do to become its own individual is to separate the mother from themselves, or themselves from the mother. Now, I'm not going to dive deep into object relations here, and it's not necessary to speculate about things like the paranoid schizoid position to understand that a child is not born with a well-developed sense of self and actually needs to learn that self and needs to learn what is an other. Unlike other animals who are born with relatively well-myelinated neurons in the brain that are efficient enough to pass basic information quickly so that they can survive, human brains are relatively immature, slow, and clunky. And because of this, our young rely on others with well-myelinated brains for survival. Myelin, for those who are struggling right now, is a covering that helps to make neurons faster, more efficient, and more isolated from their neighbors. Newborn brains are slow and inefficient and not really ready to do much more than suckle, cry, defecate, urinate, and start to learn about the world. As neurons get coated with myelin, the cortex goes from being really slow and leaky to being less slow and more specific. Sensory neurons tend to myelinate faster than motor neurons. Babies are tuned for sensing the world faster than they are tuned for acting on the world or understanding the world. Frontal lobe neurons are the last to myelinate, and these are the areas of the brain dedicated to precise control of thoughts. Your sense of self, your sense of social meaning, and your motor planning. This pattern of back-to-front brain myelination explains why speech production lags so far behind speech comprehension. Our speech sense-making areas are near the back, and they develop and myelinate their tracks, and our speech-producing areas are near the front. The motor neurons in the speech planning areas are maturing slower than those in the speech-sense-making areas. So this ever more precise deluge of sensory information is a baby's first clue as to the fact that they are a thing that can be modeled as a separate thing from the rest of the world around it. Babies also have to relatively partition off each sense from the other. A baby's brain, for instance, has to learn things like the things I see may be different than the things I am hearing. I may be able to see something that's not making a sound that I'm hearing right now. I may be able to not see the thing that's making the sound. If I see something and it is also in the direction that a sound is coming, it might be the thing making that sound. Babies are also starting to figure out things like, my finger is a part of my body, but the rattle in my hand is not a part of my body. The brain has to partition models of itself, its bodily self and its internal self from models of others, and this process continues for a very long time. Somewhere around two to three years of age, a child is more reliably able to do things like identify themselves in the mirror and start to assign states to other people. You can see this in their imaginary play, when they say things like, This doll is hungry, or mama's mad. A child may see another child reaching for food and suppose that the intention of that other child is to eat the food. This theory of mind, as it is often called, is an extension of the process of the theory of self and others. There are things that are not me, and there are things going on inside of those things that are not me. And I can predict what's going on inside those things because I know what's going on inside of me. Young children, though, even when they're able to understand that other people have their own thoughts, they often struggle to be able to predict what other people actually know because they're just assuming they know the exact same things that they know. Another skill, besides separating self from others, is separating others from others. My son's uncle, by friendship, Idris, was a very dark-skinned man from West Africa. For many months, every time my son saw another black man, he would refer to them as Uncle. The same thing happened with his uncle Johnny and Aunt Wendy, whose families had immigrated from Taiwan. Every Southeast Asian appearing person my son met for about a year or so was mistaken for either Johnny or Wendy, depending on the length of their hair. So if we can separate ourselves from others and we can separate others from others, it just follows that we can separate ourselves from ourself. There's nothing logically preventing this. There's nothing we know of in our brain that would prevent this. Our brain can do with its internal self what it does with external selves. The brain can partition itself into contexts that are relatively isolated from the other selves. Just as it can assign memories to real things that happened to me instead of imaginary thoughts that I had, it can partition memories into things that happened to alternate identities that are accessible only when that alternate self is the context in which the brain is instantiated. When the partitioning of different versions of an integrated self happens in a healthy way, then we develop what are called socially appropriate or context-appropriate selves. For example, there's myself that goes to school, myself that hangs out at home, myself that is alone when no one else is looking, and myself that tends the fire in the backyard. These selves have access to very different emotions, they act in very different ways, they have very different goals and desires, but they remain integrated. In an integrated self, each version of the self still has access to the other versions. They are partially but not fully partitioned or dissociated. They're dissociated just enough that they have efficient access to the states that they need to be in to thrive in different contexts, but not so much that one self is completely or nearly completely shut off or amnestic for the other self. Try taking a math test when you're raging mad, or when you are tired from a workout. The mind isn't operating so well in the math context, so it doesn't have easy access to the kinds of memories or algorithms that would run well in a more mathy state. But it's not completely cut off from math, and it knows how to get there, calm down, and then think its way into the math state. In normally functioning development, the adults in a child's life interact with them in a way that makes the child feel safe and separate from the adult. That adult guides the child through their different feelings, behaviors, and consequences. They allow the child freedom while setting reasonable limits. Parents talk to their children about what has happened, provide them with words and concepts to use in similar contexts. Contexts? That's a hard word. And they even point out when the child appears to be in distress, or when the child has done something that's praiseworthy. All these interactions help the child to form integrated concepts of their internal states that are relatively coherent through time and context. Despite a child's tendency to try to group others together into a single thing, for example, all the Johnnies and the Idrises in the world, feedback errors from the environment correct their models, and so they're forced to split those people apart, despite the fact that they look so much the same. The same thing happens to split the self from the rest of the world. Over time, it just makes sense that I am not a chair. However, it didn't make sense until it did. It couldn't make sense until I had had the concept that something I see across the room is not necessarily myself, even something that is looking back at me in the mirror. That thing in the mirror is not me, just a shadow of me, like a picture of a dandelion. It's not a dandelion. All these ways of giving and taking away identity are not built in. The ability to do it is built in, but the identities themselves are learned. A child's ability to learn how to group and separate is built in, but the concepts and the way that each is created is new in every individual and in every generation. That's one reason why we keep kids in school until like they're at least 18, and then we want them to go off to college, too, because it's so hard to actually make complex concepts transfer from one generation to another. The self-other concept is extraordinarily complex. The way our brains produce concepts is messy and allows for multiple dimensions that can separate very similar-looking things into very different constructs. This allows for some constructs to overlap, things like cats and dogs are both pets with fur and tails that live in the house. The way we do this makes errors in self-identity compartmentalization not only possible, but actually likely under certain conditions. I want to shatter the myth that altars represent the fragments of a shattered self. There was no self to shatter. Altars, in my opinion, are instead built by a developing brain. Some experiences cannot be logically integrated into a unified self. It's not a shattered self, it's an unintegrated self. For the time remaining, I'm going to talk about some of the ways in which the brain works that makes dissociated identities possible, and why under certain conditions, such as abuse, neglect, or chaotic environments, these identities might actually be relatively adaptive. I would never call DID a defense mechanism because I think that's kind of teleological. Like there was some goal that your mind had when it failed to integrate your personality. It assumes that DID was something the mind was trying to do. I don't believe that. I think it's a way that the mind finds itself because of what it does. And while dissociated identities might not be adaptive in general for the world, most of the time, alters arise and can be adaptive to an unpredictable and incomprehensible world. People who compete in memory competitions build something called mind palaces, where they can intentionally create all sorts of forced associations in order to remember long lists of apparently different things. Memories are not the raw sensory material of the past stored in our head. They're a reconstruction or a post diction of what the brain thinks was the most likely thing that happened or that was sensed. That's why eyewitness testimony is so problematic. Having altars supposes that each of the altars has its own unique set of memories that are hard or, in many cases, impossible to access when a person is in a different altar context. The self that we experience is merely an illusion. That the brain creates to give meaning to the present. And that goes for every self, whether it's various altars or the apparently normal self. The past is constructed by the brain to be consistent with the present in every moment. Most of the time, the past we remember seems to be a reasonable, even if flawed, representation of what actually happened. The hippocampus is the part of the brain that we most associate with memory formation and retrieval. Damage to this area can result in both posterograde and enterograde amnesia, which means we don't remember the past and can't make new memories. We used to think that the hippocampus would store a memory trace and then would call up that same trace by the same pathway as it made that trace. That's how you would access memories by just following the exact same way you made the memory. But a lab at MIT in 2017 demonstrated that the retrieval of memories actually takes a detour through a separate memory retrieval system. And this is the system that associates the memory with things like whether the experience was personal or witnessed, and what affect or emotional content was associated with that experience. What this means is that the trace that made the memory is not the trace that retrieves it. Memories are not stored precisely in the hippocampus. Instead, the hippocampus stores just enough information to be able to coax the rest of the brain into a state that simulates what happened when the memory was formed. And this simulation might even include the release of stress hormones. Modern concepts of memory formation and retrieval allow for a lot of complexity. Enough complexity that alters could plausibly exist. It does not exclude that possibility. So if memory retrieval, especially is a separate pathway from memory encoding, which has been demonstrated, this allows for certain context-dependent memories to be retrievable while others are not. We also know that as memories are retrieved, their association with certain other memories and contexts can be strengthened or weakened, depending on how they are re-experienced. If a self is a context that the brain operates in, then only memories that have been selected and associated for that context, that particular self, would be retrievable, and any new memories created in that context would be associated only with other memories that are available to that self and not to the other selves. I'm going to get a little bit more nerdy here and put this into some math language for those of you who are just itching for a computational explanation for this. I know so many of you probably are. And if you're not, maybe this will help in some vague way. Memory formation and selection could act like Bayesian model selection. So in order to prevent irrelevant memories from flooding a system with noise, just pulling up all kinds of memories all at once, something has to happen to keep that noise under control. So the brain could institute something like a softmax function that's applied during memory retrieval. And what a softmax function basically does is pushes the probability of likely related memories to near certainly related, while memories that are unlikely related are pushed to near zero probability of being related. It basically separates them out very strongly, so there's not a lot of noise and overlap. So this kind of computational trick could easily partition memories and keep them stuck in altered self-concepts. This trick can also help explain how alters can emerge in the first place. So as the brain is combing through its memory space, some of the memories have not been fully partitioned by the softmax function. And if one of those memories then provides access to a memory that has been more fully partitioned, then applying an iterative softmax function to continue memory recall could push the brain into what could be called the different alter state. I'm going to summarize that by saying if there's a little bit of overlap left, then some memories can be associated with other memories in the altered state. But once that other memory from the altered state is called up, it pushes the rest of the brain into the alter state. Because memories associated with that one are more strongly the memories from that alter state, and the other memories are not. This could be like a trigger. And mathematically, it could make perfect sense. Basically, you end up in something called a different attractor state. Now I've just used a lot of words that might be confusing, like Bayesian, softmax, and attractor states. But to summarize, the most recent evidence for how the brain works suggests that the process of memory retrieval is fundamentally different from the process of memory formation. Memories are more or less accessible, depending on how well they relate to other memories in the brain, while you are recalling memories. In order to not get confused, the brain strengthens the association between what it thinks are related memories while at the same time pruning away the association between unrelated memories. This means that there could be islands of memories that are relatively partitioned off from each other, and any incompletely partitioned memory might act as a bridge between these two islands. Now I want to pause here and repeat that my goal is not to propose that I or anyone else has completely figured out how DID works, but only to demonstrate that our neurobiology logically allows for it under certain circumstances. The conditions that most reliably increase the chance of someone developing alters are experiences that seem irreconcilable, unable to be related to our other experiences. They don't seem to fit. For example, when a caregiver is both caring and then abusive, alternately attentive and neglectful. Or when there's chaotic structure and very little corrective feedback that it can allow the child to develop an integrated self from this larger set of apparently different experiences that the brain can't make sense of. DID is not a fractured self, but an incoherent self. It results when the mind, applying its ability to rate relate things to each other, cannot find ways to reconcile some memories with others, and so various groups or islands of memories form, and their associated states coalesce. And these islands are your alters. We know that in traumatic experiences, some people experience amnesia for the events of the trauma or for events surrounding the trauma. Memories are not associated with other memories in a structured way, often during traumatic events. So the memory encoding doesn't really allow for that memory to be recalled faithfully. Occasionally, some signal in the brain might create a bridge to that memory and that the brain enters the traumatic state again. But getting back to dissociative identity disorder, when a brain is in any one particular state, whether it is a relatively focused state needed to complete things like an exam or a more intense state needed to survive the next few minutes, any information in the brain that is not relevant to the context is actively attenuated or gated. I can't be pondering the shape of the clouds while trying to figure out the impedance in an electrical circuit. And I can't be considering that pain in my leg when I need to be running away from my abuser. Another thing that can happen in context-dependent states is that even if the brain does not have access to its apparently normal memories, it has to be able to access something and make decisions. Therefore, the brain starts to make decisions that are based on far less information. It has to. The brain will make a decision or just freeze up. If something really seems to not make much sense, or there are holes in the sense making, then the brain has another trick that it uses. This trick is called confabulation. Confabulation is the result of the brain making inferences about what must be true, and since there's little information to challenge those inferences, they're adopted, and made-up experiences can be experienced as highly certain facts. The brain must invent a reality that makes sense or it will experience catastrophic failure. This confabulated experience can appear as a real and logical experience. It's as real and logical to the brain as any reality that it is actually experienced. All of us have confabulated memories. How much of our memory is confabulated versus how much is reality-based can vary greatly between individuals. When DID 1.0 was first defined as multiple personality disorder, a debate raged about whether or not most of the explosion of DID diagnoses were in fact different altars or if they were the result of patients who are highly suggestible, being convinced unwittingly by well-meaning therapists that they have separate personalities inside of them. Now, it's likely that some subset of patients diagnosed with DID do not actually have DID. But the evidence that altars are implanted by therapists, that they are constructs the therapists put inside of patients, just doesn't really hold water. Most therapists are not hypnotizing their patients and then trying to implant false memories. And most people are not that suggestible. So DID didn't arise because of a bunch of therapists brainwashing their patients. More and more, evidence shows there's very strong relationships between dissociative identity and early childhood trauma and abuse and neglect. There's almost no evidence that people are as suggestible as it would require for them to be able to adopt 12 or 16 different personalities and switch between those altars. The juice just wouldn't be worth the squeeze. However, because DID patients have their generally present, social-facing, apparently normal self as their context most of the time, the other altars can suffer from sort of a lack of learning or a lack of information. And this can result in them, the other altars, being more susceptible to confabulating things. They could confabulate their past. They might even confabulate reasons for the things that they're doing in the moment. These constructed identities are necessarily influenced by the social environment that they're formed in, but it doesn't mean that they were created by it. For example, one of the first likely documented cases of dissociative identity disorder I talked about in the 16th century, a nun named Jean Fairy. Now she was treated with what they knew at the time, exorcism. Her altars included benevolent protectors like Mary Magdalene, and then misguided protectors like Kornao and Garga, who were present when she was engaged in behaviors like self-strangulation or disordered eating. Altars in Fairy's context were considered devils or demons who had possessed her body. And the protective spirit, Mary Magdalene, may have arrived to exercise those demons. Her altars took those forms because those forms were available and could be confabulated into place. Religion did not cause her altars, but it did give a social context for those altars to manifest themselves. Because altars are not, in fact, demons possessing the body, they will necessarily follow the cultural scripts of the time, and because some cultures will have less access to actual experience than others, some will have to rely more on confabulation to understand the world when they're instantiated. So, because of this, some altars might seem far more organized and coherent than others. People also often describe altars as having a purpose, such as a caretaker, protector, or persecutor. There's altars that are referred to as the emotional part, and often these altars are described as holding the vast majority of traumatic memories. Others might claim to have access to various altars that act either as like a gatekeeper who can influence when there's a switch between altars, or as a guide who can help explain the other altars. So it's possible that one altar might have some limited access to the memories of the others. How one memory tracing follows another and whether that leads to a full switch or an open window is not well understood. Whether the apparently normal part is aware that altars exist can have a large impact on how the disorder progresses. Some people with DID report having thoughts they don't feel belong to them or hear what appear to be voices in their head, but these are not psychotic voices. This is all to say that if I could go back in time to my previous self, I would tell myself that the reason that I did not think that DID was a probable diagnosis is because I didn't know enough about how the brain works to even have an informed opinion about it. I have since changed my mind. And in doing that, I've changed myself. I want to end by just wildly speculating. Not that I haven't done some speculation already, but I think it's good to imagine things and then share them with others. Something I've been thinking about recently is whether the formation of altars can be seen as part of normal development, not just abnormal development. DID, as I presented it previously, is conceived of as a failure of self-integration. But what is it that the brain is integrating? I'm starting to wonder if during development children actually naturally form many different islands or memories attached to various selves, and that this happens normally to begin with. In effect, there are altars inside of all of us that persist for some period of time, but then these altars are slowly integrated as connections are made between them and the different islands kind of coalesce into larger islands and cease to be isolated. In this framework, we kind of all start out with dissociated identities in our normal development and then integrate these identities. It could be that trauma, abuse, neglect, or unpredictable environments are not creating these islands, but are just pushing the islands further apart, interrupting their integration, such that they then start to form their own relatively independent attractor states that have less and less access to the others. In some instances, these altars might just fade away with time, especially during normal development. If one part of the brain hasn't coalesced with the other, it could just fade away. But in the event that there is trauma or more severe or chronic neglect, altars could be instantiated and then reinforced enough that they don't just fade away. Instead, they persist as deep well attractors. Thank you for taking some time to listen to me ramble and to speculate. If you have any questions or want to send me feedback, then you can email me at feedback at psyactic.com. You can also go to psydactic.com and fill out a form there. I would love to continue this conversation. I am Dr. O'Leary, and this has been an episode of Psydactic.