Good question! I think there are two different steps here. (The following is a bit oversimplified.)
Step 1 is auto-associative recall in the hippocampus. Neurons all over the cortex directly or indirectly activate neurons in the hippocampus. And then if something is happening in any part of the cortex that partially matches some old memory, the whole old memory can autocomplete within the hippocampus.
Step 2 is: that core of an old memory has lots of (direct and indirect) associations all around the cortex / global workspace. Like, if it’s a visual memory, the neurons in the hippocampus will send signals that the relevant visual region should go into the corresponding activation state.
Loosely speaking, there’s local “attractor dynamics” within the hippocampus (Step 1) and simultaneously there’s larger-scale “attractor dynamics” around much of the cortex (Step 2).
I keep using the example of having Tune A in your head (or actually hearing it) and then seeing something that would normally trigger a memory of a different Tune B (assuming no special training :-P). In this situation, I think Step 1 goes through, i.e. part or all of the hippocampus does in fact transiently get into a state that represents Tune B. But then Step 2 is blocked. Instead of there being a large-scale “attractor” around the cortex that grows out of the “seed” at the hippocampus, instead that Tune B attractor fails to spread, instead getting blocked by the existing Tune A attractor, which pushes back and wins.
More generally, we can ask which attractor will “win” in a fight between two partially-formed mutually-incompatible atteractors. It’s complicated, but attention has a lot to do with it. If you’re holding rapt attention towards what you’re looking at right now, then you’re less likely to spontaneously visually imagine the appearance of some random thing you remember from earlier in life. Whereas if you’re not particularly paying attention to anything, then there’s nothing stopping a visual memory from popping up.
So back to your question, when you recall the childhood treehouse, I claim there’s at least a brief moment when you stop paying attention to current visual inputs and instead use (part of) your visual processing capacity to imagine the treehouse. Then you flip back to what you’re looking at. It all happens very fast—a fraction of a second. So I think there’s an attentional blip associated with briefly thinking about the treehouse, but we don’t really notice, just as people are surprised to learn about change blindness and the selective attention task and so on.
So then you can ask: if I can transiently stop paying attention to visual input in order to recall what my old treehouse looked like, why can’t the DID patient likewise transiently stop paying attention to the concept of Alter A in order to imagine the concept of Alter B? (…at least in some cases.) I think that the answer again centers around attention. I think there’s a tendency to hold attention on the homunculus concept, whatever it is, not enough to make it the center of attention most of the time, but enough to prevent it from disappearing entirely.
In this respect, I think the homunculus concept is less like something I happen to be looking at, and more like “the idea that I have an itchy knee” or “the idea that I’m feeling anxious”. Again, these don’t need to be the center of attention. But they do seem to reliably benefit from a steady drip of low-level attention, such that they tend to be kinda looming in the background, whatever you’re thinking about.
Why does the homunculus have that special property, of holding low-level attention in a way that most concepts don’t? I think it’s some combination of (1) In general, the homunculus is associated with mild physiological arousal (since one of its defining properties is “surprisingness”, see §3.3.2), which makes it inherently mildly attention-grabbing; (2) In DID in particular, there’s an association between the alters and moods, and moods like anxiety and anger and sadness have their own involuntary attention mechanism.
There’s a bit more about some of this in the next post, actually. :)
(Oh, here’s another possibility, I suppose. I think the reason a tune in your head tends not to be interrupted by a remembered tune, is that we have a general habit of holding continuous voluntary attention on tunes because that’s the only way to not “lose our place”. This also applies to anything else where there’s a risk of “losing one’s train of thought”: attention is unusually strong, because you’re trying to block random memories or daydreams from butting in. Maybe there’s something analogous, where we have a self-reflective “train of thought” related to “what the homunculus is doing”, related to the narrative of our day and lives and so on, and this is sufficiently useful and motivating that we learn to keep voluntary low-level attention on the homunculus almost always. …I think I prefer the other theories above though, related to involuntary attention rather than voluntary attention.)
Good question! I think there are two different steps here. (The following is a bit oversimplified.)
Step 1 is auto-associative recall in the hippocampus. Neurons all over the cortex directly or indirectly activate neurons in the hippocampus. And then if something is happening in any part of the cortex that partially matches some old memory, the whole old memory can autocomplete within the hippocampus.
Step 2 is: that core of an old memory has lots of (direct and indirect) associations all around the cortex / global workspace. Like, if it’s a visual memory, the neurons in the hippocampus will send signals that the relevant visual region should go into the corresponding activation state.
Loosely speaking, there’s local “attractor dynamics” within the hippocampus (Step 1) and simultaneously there’s larger-scale “attractor dynamics” around much of the cortex (Step 2).
I keep using the example of having Tune A in your head (or actually hearing it) and then seeing something that would normally trigger a memory of a different Tune B (assuming no special training :-P). In this situation, I think Step 1 goes through, i.e. part or all of the hippocampus does in fact transiently get into a state that represents Tune B. But then Step 2 is blocked. Instead of there being a large-scale “attractor” around the cortex that grows out of the “seed” at the hippocampus, instead that Tune B attractor fails to spread, instead getting blocked by the existing Tune A attractor, which pushes back and wins.
More generally, we can ask which attractor will “win” in a fight between two partially-formed mutually-incompatible atteractors. It’s complicated, but attention has a lot to do with it. If you’re holding rapt attention towards what you’re looking at right now, then you’re less likely to spontaneously visually imagine the appearance of some random thing you remember from earlier in life. Whereas if you’re not particularly paying attention to anything, then there’s nothing stopping a visual memory from popping up.
So back to your question, when you recall the childhood treehouse, I claim there’s at least a brief moment when you stop paying attention to current visual inputs and instead use (part of) your visual processing capacity to imagine the treehouse. Then you flip back to what you’re looking at. It all happens very fast—a fraction of a second. So I think there’s an attentional blip associated with briefly thinking about the treehouse, but we don’t really notice, just as people are surprised to learn about change blindness and the selective attention task and so on.
So then you can ask: if I can transiently stop paying attention to visual input in order to recall what my old treehouse looked like, why can’t the DID patient likewise transiently stop paying attention to the concept of Alter A in order to imagine the concept of Alter B? (…at least in some cases.) I think that the answer again centers around attention. I think there’s a tendency to hold attention on the homunculus concept, whatever it is, not enough to make it the center of attention most of the time, but enough to prevent it from disappearing entirely.
In this respect, I think the homunculus concept is less like something I happen to be looking at, and more like “the idea that I have an itchy knee” or “the idea that I’m feeling anxious”. Again, these don’t need to be the center of attention. But they do seem to reliably benefit from a steady drip of low-level attention, such that they tend to be kinda looming in the background, whatever you’re thinking about.
Why does the homunculus have that special property, of holding low-level attention in a way that most concepts don’t? I think it’s some combination of (1) In general, the homunculus is associated with mild physiological arousal (since one of its defining properties is “surprisingness”, see §3.3.2), which makes it inherently mildly attention-grabbing; (2) In DID in particular, there’s an association between the alters and moods, and moods like anxiety and anger and sadness have their own involuntary attention mechanism.
There’s a bit more about some of this in the next post, actually. :)
(Oh, here’s another possibility, I suppose. I think the reason a tune in your head tends not to be interrupted by a remembered tune, is that we have a general habit of holding continuous voluntary attention on tunes because that’s the only way to not “lose our place”. This also applies to anything else where there’s a risk of “losing one’s train of thought”: attention is unusually strong, because you’re trying to block random memories or daydreams from butting in. Maybe there’s something analogous, where we have a self-reflective “train of thought” related to “what the homunculus is doing”, related to the narrative of our day and lives and so on, and this is sufficiently useful and motivating that we learn to keep voluntary low-level attention on the homunculus almost always. …I think I prefer the other theories above though, related to involuntary attention rather than voluntary attention.)