Friday, September 08, 2006

We tell ourselves stories   posted by Coffee Mug @ 9/08/2006 09:14:00 AM

The September issue of Hippocampus is a special on Hippocampal Place Fields and Episodic Memory. Episodic memory is the type of memory that you immediately think of if you are not a memory researcher. It is what Proust is doing while Remembrancing (which is part of why he is such a favorite quotee among the memory community). EM is supposed to have key elements that distinguish it from other memory types, one of which is this experience of "mental time travel". This is an idea almost entirely based in introspection. It is not clear how to ascertain if a patient is replaying the experience subjectively or merely retrieving facts without the VR-like aspect. The subjective nature of this type of memory retrieval makes it well-nigh impossible to study in rodents. But it may be that people who are mostly concerned with the mental time travel issue are hardly interested in memory, per se, but are rather interested in our ability to "fill in the gaps" and create continuous narratives. From the second article of the issue by Ferbinteanu et al.:
Furthermore, though intuition also suggests that our memories are veridical-an accurate reproduction of past events-empirical data indicate that autobiographical memories are in fact reconstructed by active processes sensitive to systematic errors based upon inattention, suggestion, expectancy, and familiar cognitive scripts (Schacter, 1999; e.g., Conway, 2001b). Even completely false memories are acquired easily (Loftus, 1997, 2004) and activate the same neural network involved in true memories (Okado and Stark, 2005). These memory distortions show that rather than "traveling down the memory lane" to re-experience past events, memories for episodes are reconstructed representations based on fragmentary data fit together using heuristics (Schacter, 1999; Conway and Pleydell-Pearce, 2000).
So, you don't really remember. You confabulate based on the information you have at hand, the "fragmentary data". I have gotten so paranoid about confabulation that I will hardly ever state one of my memories of a past event as fact. I would be a horrible murder-trial witness. I notice errors all the time among friends of mine that are better story-tellers. We will be reminiscing about some event 4 or 5 years ago and place some individual there that we didn't even know yet. Everyone knows we have memory errors. I wonder if better story-tellers are better at "smoothing the curve". We have a bunch of data points stored up, but it may take our narrative-building ability (which may be uniquely human) to experience "mental time travel".

This habit humans have of telling ourselves stories seems pervasive. I am reminded of the hyperactive agency detection device, proposed as an explanation for our need to make gods. Also, Gazzaniga's experiments with split-brain patients:
Studies on split-brain patients have dominated Dr. Gazzaniga's work ever since. In the 1970's, he and his colleagues reported that the left hemisphere acts as an interpreter, creating theories to makes sense of a person's experiences.

Their first clue came from an experiment Dr. Gazzaniga carried out with Dr. Joseph LeDoux, now at New York University. A patient called P.S. was shown a picture, and was then asked to choose a related image from a set of other pictures. What P.S. didn't know was that he was being shown a different image in each eye.

Dr. Gazzaniga and Dr. LeDoux showed P.S. a picture of a chicken claw in his right eye and a snow-covered house in the left eye. P.S. pointed to a chicken with his right hand and a snow shovel with his left.

''I'll never forget the day we got around to asking P.S., 'Why did you do that?''' said Dr. Gazzaniga. ''He said, 'The chicken claw goes with the chicken.' That's all the left hemisphere saw. And then he looks at the shovel and said, 'The reason you need a shovel is to clean out the chicken shed.'''

Dr. Gazzaniga hypothesized that P.S.'s left hemisphere made up a story to explain his actions, based on the limited information it received. Dr. Gazzaniga and his colleagues have carried out the same experiment hundreds of times since, and the left hemisphere has consistently acted this way.

''The interpreter tells the story line of a person,'' Dr. Gazzaniga said. ''It's collecting all the information that is in all these separate systems that are distributed through the brain.'' While the story feels like an unfiltered picture of reality, it's just a quickly-thrown-together narrative.
The point of the Hippocampus article is that we can ignore this business for now, while we study how the hippocampus brings together the information that we actually have remembered rather than confabluated. This process is amenable to animal studies, which is great because we are getting really good at multi-electrode recording in the rat hippocampus and gaining lots of insight into place cells, which may provide the spatial context for remembered events to happen in. Which brings up an interesting question: Do you ever remember an event without remembering where you were during that event? Or does the term "event" imply a spatial setting?