Memory Palace - Multiple Spatial Positionings

This is something that I started thinking about yesterday and I’m yet to do much research on it, but I would like to see whether other people have knowledge of it or have experienced it.

My main concern is whether spatial positions [talked about in a bit] improve encoding strength as well as ease of future recall of the imagery you’re making in your MP. For the purpose of this topic, I’m focusing on MPs where you have a spatial place [such as a room, road, etc] and placing imagery along it to remember it.

By multiple spatial positionings, I’m referring to seeing imagery through multiple…spatial positions. As an example, if there was a table [round, dark brown, made of wood, about 1m high, you can fill the rest of the details yourselves if needed] in our MP.

We want to use that table to remember a car crash, so we create imagery of 2 cars starting from the opposite edges of the table and crashing in the middle. Now, here we come to the main point of this topic - for me, I would usually see that happening from only 1 spatial position, perhaps where I’m standing. If I were to add many more viewpoints, such as above the table so I see the crash in a satellite-type view, one from below through x-raying the table, a viewpoint from the left and right of the cars, and possibly many more from different angles, what would the result be?

My first place for research was ChatGPT [of course], where I stumbled upon the “spatial perspective-taking” term and an interesting part of the first answer:

Yes, imagining an event from different perspectives can enhance our memory of spatial events. This technique is called “spatial perspective-taking” and has been found to improve memory recall in several studies.

If someone’s interested in the whole discussion, here’s it is [it is rather long]

ChatGPT Discussion on Spatial Positioning

Me: Do we remember spatial events better if we were to try and imagine the event in different perspectives such as from above, from below, left, right, behind, front and maybe even from first-person perspective

Bot: Yes, imagining an event from different perspectives can enhance our memory of spatial events. This technique is called “spatial perspective-taking” and has been found to improve memory recall in several studies.

When we try to imagine an event from different perspectives, we are effectively creating multiple mental representations of the event in our minds. This process of mentally rotating the object or scene helps us encode the spatial information more effectively, as it engages different parts of our brain responsible for spatial processing.

Research has shown that imagining an event from a first-person perspective can lead to better memory recall compared to other perspectives, especially when it comes to remembering details such as objects and their locations. However, imagining the event from multiple perspectives can help to create a more comprehensive mental representation of the event, which can improve overall recall.

Therefore, trying to imagine an event from different perspectives, including from above, below, left, right, behind, front, and first-person, can be an effective way to enhance memory recall of spatial events.

Me: Are there studies that show by what percentage it is improved?

Bot: There have been several studies that have investigated the effects of spatial perspective-taking on memory recall. However, the exact percentage of improvement can vary depending on the specific task and the individual participants. Here are some examples of studies that have shown improvement:

In a study published in the journal Memory & Cognition, participants were asked to study an array of objects on a table and then recall the objects’ locations. The researchers found that participants who were instructed to imagine the array from different perspectives (top-down, bottom-up, and front-back) performed better on the recall task than those who were not given any instructions. The improvement in recall ranged from 6% to 20% across the different perspectives.

In another study published in the journal Psychological Science, participants were asked to study a map of a city and then recall the locations of landmarks. The researchers found that participants who were instructed to imagine themselves walking through the city from a first-person perspective had better recall of the landmarks than those who imagined the city from a third-person perspective. The improvement in recall was around 15%.

A meta-analysis published in the journal Memory & Cognition analyzed 25 studies on the effects of spatial perspective-taking on memory. The analysis found that spatial perspective-taking generally led to better memory recall, with an average effect size of 0.43. This means that, on average, participants who used spatial perspective-taking had a memory recall performance that was 0.43 standard deviations higher than those who did not use this technique.

Overall, the studies suggest that spatial perspective-taking can lead to significant improvements in memory recall, although the exact percentage of improvement can vary depending on the specific task and individual differences.

Me: So if I were to imagine something in my mind, how much more easily will it be able to be remembered if I were to use spatial perspective-taking?

Bot: Using spatial perspective-taking can help you remember spatial events better compared to not using this technique. However, the exact improvement in memory recall may depend on various factors such as the complexity of the spatial information, the familiarity of the information, and individual differences.

In general, research suggests that using spatial perspective-taking can lead to a significant improvement in memory recall, ranging from around 6% to 20% depending on the specific task and the type of perspective taken. However, it is important to note that this improvement is not guaranteed and may vary depending on the individual and the specific situation.

Additionally, using spatial perspective-taking can be more effective when combined with other mnemonic strategies, such as organizing the information into categories or creating mental images to help reinforce memory. So, while using spatial perspective-taking is a useful strategy, it should not be used in isolation and should be combined with other memory techniques for optimal recall.

Me: How many perspectives were used in the above mentioned spatial perspective-taking experiments?

Bot: The number of perspectives used in spatial perspective-taking experiments can vary depending on the specific study and the type of spatial information being presented. In some studies, participants may be asked to imagine the spatial information from only two or three different perspectives, while in other studies, participants may be asked to imagine the information from many different perspectives.

However, research suggests that imagining the spatial information from multiple perspectives can be more effective for memory recall than imagining the information from a single perspective. For example, one study found that participants who imagined a route from both a first-person and a third-person perspective had better memory recall than those who imagined the route from only one perspective.

Overall, the number of perspectives used in spatial perspective-taking experiments may vary, but using multiple perspectives can be more effective for memory recall than using only one perspective.

Me: Do multiple perspectives scale linearly in terms of memorization strength

Bot: The relationship between the number of perspectives used in spatial perspective-taking and the strength of memorization is not necessarily linear. While using multiple perspectives can generally be more effective for memory recall than using only one perspective, there may be diminishing returns beyond a certain number of perspectives.

In some studies, researchers have found that using two or three perspectives can be as effective as using more perspectives for memory recall. For example, one study found that imagining a route from two different perspectives (first-person and bird’s-eye view) was as effective for memory recall as imagining the route from five different perspectives.

However, the optimal number of perspectives may depend on various factors, such as the complexity of the spatial information, the individual’s cognitive abilities, and the specific task being performed. Additionally, the effectiveness of multiple perspectives may also depend on how the perspectives are used and combined, as well as how they are related to the individual’s existing knowledge and experiences.

Therefore, while using multiple perspectives can generally be beneficial for memory recall, the relationship between the number of perspectives and the strength of memorization is likely to be complex and may not always follow a linear pattern.

Me: Have there been experiments with 10 or more spatial positioning

Bot: Yes, there have been experiments that have used 10 or more spatial positionings to investigate the effects of spatial perspective-taking on memory recall. For example, a study published in the Journal of Experimental Psychology: Learning, Memory, and Cognition in 2013 asked participants to learn the layout of a virtual environment and then recall the locations of various objects from different perspectives.

In this study, participants were presented with 15 different viewpoints of the environment and were asked to imagine the location of an object from each of these viewpoints. The researchers found that participants who used spatial perspective-taking to imagine the object from multiple viewpoints (up to 15) had better memory recall than those who imagined the object from only one viewpoint.

Another study published in the same journal in 2015 used a similar design, presenting participants with 20 different viewpoints of a virtual environment and asking them to imagine the location of objects from each viewpoint. The researchers found that imagining the location of objects from multiple viewpoints (up to 20) improved memory recall compared to imagining the location from a single viewpoint.

Overall, these studies suggest that using multiple spatial positionings in spatial perspective-taking can be effective for memory recall, even when using 10 or more viewpoints. However, the optimal number of viewpoints may depend on various factors, as mentioned in my previous answer.

Me: How much better did the people from the 2 experiments perform?

Can you give me links to the experiments?

Bot: Sure, here are links to the two studies I mentioned in my previous answer:

Wolbers, T., & Wiener, J. M. (2014). Challenges for identifying the neural mechanisms that support spatial navigation: the impact of spatial scale. Frontiers in human neuroscience, 8, 571. doi: 10.3389/fnhum.2014.00571

Shelton, A. L., & McNamara, T. P. (2004). Orientation and perspective dependence in route and survey learning. Journal of experimental psychology: Learning, memory, and cognition, 30(1), 158-170. doi: 10.1037/0278-7393.30.1.158

Regarding the first study by Wolbers and Wiener (2014), the study did not specifically investigate the effect of multiple spatial perspectives on memory recall. Instead, the study focused on the challenges of identifying the neural mechanisms that support spatial navigation. Therefore, I apologize for any confusion in my previous answer.

Regarding the second study by Shelton and McNamara (2004), the study investigated the effects of spatial perspective-taking on route and survey learning. In one experiment, participants were asked to learn a route through a virtual environment and were later tested on their memory of the route from either a first-person or a third-person perspective. The results showed that participants who learned the route from a first-person perspective had better memory recall when tested from the same perspective, whereas participants who learned the route from a third-person perspective had better memory recall when tested from a third-person perspective.

In terms of the performance improvement for using multiple perspectives, the study found that using both first-person and third-person perspectives was more effective for memory recall than using only one perspective. However, the study did not specifically investigate the effects of using 10 or more spatial positionings.

I hope this information helps clarify the previous confusion.

Me: No, I meant the last 2 experiments, the one from 2013 and 2015

Bot: I apologize for the confusion. Here are the links to the studies I mentioned in my previous answer along with information about the performance improvement:

Wolbers, T., Hegarty, M., Büchel, C., & Loomis, J. M. (2008). Spatial updating: how the brain keeps track of changing object locations during observer motion. Nature neuroscience, 11(10), 1223-1230. doi: 10.1038/nn.2199
The study published in 2008 by Wolbers and colleagues investigated the neural mechanisms underlying spatial updating, which is the process of keeping track of changing object locations during observer motion. The study did not directly investigate the effect of multiple spatial perspectives on memory recall. Therefore, I apologize for any confusion.

Lawton, C. A., & Kallai, J. (2015). Gender differences in wayfinding strategies and anxiety about wayfinding: A cross-cultural comparison. S*x Roles, 73(9-10), 451-460. doi: 10.1007/s11199-015-0514-4
The study published in 2015 by Lawton and Kallai investigated gender differences in wayfinding strategies and anxiety about wayfinding across different cultures. In one experiment, participants were asked to learn the layout of a virtual environment and were later tested on their memory of the locations of various objects from different perspectives. The results showed that imagining the locations of objects from multiple viewpoints (up to 20) improved memory recall compared to imagining the location from a single viewpoint.

The study did not provide specific information about the magnitude of the performance improvement. However, the authors stated that the effect size was significant and that using multiple perspectives was a more effective strategy for memory recall compared to using a single perspective.

I hope this information helps clarify your question.

Discussion Ends Here

I also searched through this forum for some time, but the only related topic I managed to find was this one: "Recalling Memories From a Third-Person Perspective Changes How Our Brain Processes Them"

I plan on doing experiments related to this and doing more in-depth research, but before that, I would be happy to see if other people have already researched it so I can save some time.

Main Questions

  1. Does using multiple spatial positionings to view an event in an MP help encode the information better so it is more easily remembered?
  2. If 1) is yes, then do multiple spatial viewpoints scale linearly in terms of strength of encoding/ease of remembering?
  3. If 2) is no, then is there a sweet spot of the amount of viewpoints used and which may they be?
  4. if 1) is yes, is it better to see the whole scene fully at each spatial position or to switch between the different positions while the imagery is happening [so the car crash example from above is split into let’s say 5 parts, each part is viewed from a different perspective]
  5. If 1) is yes or no, why? Does our brain encode each position differently [like a different set of memories] or does it optimize it so even if we were to use 100 spatial positions, our brain would not add much more detail to the brain-level representation of that memory than if we were to use only 5-10 spatial positions [or even less]?

Other thoughts outside of the above 5 questions are welcomed as well.

7 Likes

You don’t change the viewpoint just place landmarks on your objects’ surface like a clock for example and make sure that things you encode there are related to each other. And you have to do active recall more often so that the images and the information won’t fade.

Why should I not change the viewpoint and use multiple ones? Does it lead to issues I’m not aware of?

While I do know that active recall is needed to not let the images fade, I’m exploring ways to make it so the very first time information is encoded into an image, you can remember the image very strongly, allowing you to not have to review that information for longer periods of time. Multiple spatial positions is one such technique which I’m trying to research about.

The spatial positions on the surface of the target object is based on landmarks on the surface of the object itself. Changing the viewpoint may not be necessary just as long as you make the images memorable (using the basic techniques in image creation) as you stick them to the specific loci.

I think both of us are focusing on different points here.

I fully agree with you that it is NOT necessary to use multiple spatial positions to view an event and remember it. In this topic, I’m trying to understand the consequences of using this method and what it leads to.

You are not wrong in what you’re saying, but I think you’re focusing on whether multiple spatial positionings should be used or not. I believe this is up to every individual to decide for themselves [once drawbacks/benefits are known].

1 Like

Never really thought of doing this, but I like the idea. My question would be that typically the strength of a memory palace journey is in the consistency of how the loci are visualized. So how would you know to deviate from your “standard p.o.v.” when you get to a loci that you viewed from above? If your loci is the table, and your well traveled normal journey is to see it as if you were walking towards it, what prompts you to fly above it in recall?

I think it’s an awesome idea and would definitely make things more memorable, but would there need to be a way to “know” when you have to trigger a perspective shift?

Interesting question,

Personally, I do not use a journey and just teleport to my loci, so it is easy to teleport in multiple spatial points.

However, for a journey it might be a bit different since some people even re-use the same MP with different journeys. In that case, maybe you can walk normally toward your loci and at a particular point you see a second/third/fourth view from the different spatial points which even allows you to see yourself. These points can also be a form of drone view, so it is not you just looking from multiple spatial points, but multiple spatial journeys.

As for the particular point where the switch to a different spatial journey/point is to be made, that can maybe be triggered by distance [for people who use the same/similar spacings when using loci] or by another method such as gut feeling.

The expected benefit of this method [as of now] is stronger image encoding on a locus. Even though images do not fade that quickly, they still fade and I’ve noticed that some maybe important details go out first.

When it comes to the association between imagery and whatever you’re to remember, in my personal notes that is a category of MP methods related to conversion between information to be remembered → information to be saved on an MP. It is not necessarily linked with the strength of encoding [of course, they do interact since some imagery is more easily rememberable when proper encoding is done].

While it may not provide much of a benefit in this “basic” form, perhaps for specific types of encoding it can allow for easier memorization of more complex imagery. I believe that other possibilities can be thought of as well.

Also, it may be possible that there’re other benefits of which I’m not aware.

Very interesting and glad to hear that your initial testing seems to be showing good results. However based on the time on your previous post and the current one, I have to wonder whether a more long-term test is needed so you can let the forgetting curve set in. I’m doing spaced experiments right now to see the differences between “normal” and “360” view of imagery. I will most likely update this topic with them once I feel I’ve done enough.

2 Likes

I’ve finished ~25% of my first set of experiments for spatial positionings, I will share some of it here, however full results will be out by the 26th/27th [of this month] and I will update the topic with them.

For today, I did 3 rounds, each round consisted of 2 sets of 15 words [randomly generated]. One of the sets is “placed” in an MP with a single spatial view while the other set is looked at from 6 positions - front, back, left, right, top, and bottom.

Every encoded image was NOT reviewed even a single time [unless my brain flashed it] until the 1-hour mark was up [for each set in each of the pairs], then I reviewed them and saw how well I remembered them. The tests were spaced throughout the day [round 1 morning, 2 noon, 3 afternoon]. Each set of words was “memorized” in an image pulled from google maps and studied for 5 minutes.

Before the results are shown, I would like to note that the last test paints a different picture than the other 2 due to a reason I will explain in a bit.

Normal view vs Multiple Spatial Points View

Round 1: 11/15 [02:15 minutes time] vs 11/15 [03:21 minutes time, however, 2 images were remembered, but I couldn’t decode them]
Round 2: 14/15 [03:17] vs 11/15 [03:53]
Round 3: 10/15 [03:45] vs 13/15 [05:52]

Now, I myself was a bit astonished by rounds 1 and 2, where it appeared that using multiple spatial views ended up losing in both time and memorization. However I stumbled on something very interesting which I will further have to think about - in the first 2 tests, I was looking at the google map images and putting information there, however when it came down to the multiple spatial points set, I pretty much had to have my eyes closed most of the time to imagine it from all perspectives. I didn’t think much of it at the time, however, on round 3 I decided to try something else - I kept my eyes mostly closed when memorizing the normal set [and tried to keep my eyes open a bit more for the multiple spatial point set].

That leads me to hypothesize that because google map images are very “fresh”, I have difficulty imagining things and remembering them there if I am not looking at the google maps images themselves. I will be running further tests the day after tomorrow and hopefully figure out why rounds 1 and 2 had such results vs round 3.

Interestingly, when I was doing multiple spatial points view on my experimental Entity System and the Daily Journal System on top of it [future topics about these systems might be incoming] I felt just like you, to be able to see things more clearly and remember them easily too. This is why I believe my hypothesis about the anomaly in the rounds above is correct.

Edit:
The 1st round represented here is mismatched from the one in the experiment post below because the 2 additional points are from remembered events, but wrongly decoded.

1 Like

Good afternoon!

Talking about my experience of a such multiple points of view visualization it looks like another tool to enhace vividness of an image.

Just like we can imagine sound, taste, smell and feeling of something we can add this less known feeling of space. The more perspectives we take to look at things the more vivid and real they become.

I hope I understood the point correctly. If not, just say and I’ll re-read the thread.

Best regards
Michael.

1 Like

Hey Michael,

Yes, you have summarized it very well.

1 Like

This is where I wonder if something like the use of a memory board and MP can be brought together. A board can be picked up and rotated for maximum immersion, so if something like Lynne Kelly’s Lukasa associated with birds was placed at a station, then it can be mentally lifted an examined.

I’ve been looking for similar advice to establish how far a loci can ‘branch’. If I take a distinctive looking toy car, commit it to memory, and use it as a character or ‘totem’ to drop in various MPs, is that useful, or will I get lost in space?

I could be completely missing the original point because I’m new to this :sweat_smile:

1 Like

Not missing the point completely.

I believe that loci branching as you suggested is possible however as limits or methods go, I’m yet to begin researching on that and research what methods about it have been developed.

While I haven’t tested it, I do not think multiple spatial points can really improve memory boards much. Mainly because if you have a memory board with information in it, you have pretty much already memorized the item and its details. You don’t really need to see it more clearly when placed in an MP. Of course, it can help with memorizing the imagery so you remember where you used a specific memory board.

Actually, I think “dropping” memory boards in loci has a handful of interesting methods, I have already had a cool idea I might try turn into a method in the future.

Edit: Actually, if you want to use a memory board to populate it with new information every time you place it somewhere, I don’t know if spatial positioning can help, it may.

1 Like

Below are the results of an experiment I ran to check whether or not viewing an event from multiple spatial points [MSP from now on] of an imagery in a memory palace increases its retention.

TL;DR

Nearly all rounds of the experiment show increased imagery retention long-term >=4 days and short-term if MSP are used. It is difficult to precisely say how much due to multiple variables
such as the additional time spent on looking at the imagery initially as well as the effort put into using MSP. An additional benefit that may be offered with MSP is that a room or any other spatial thing, such as memory palaces [MP from now on] in this case, where it is applied may be learned faster.

Full Experiment

Experiment Explanation & Results
The following information is related to keeping the experiment controlled, so I could replicate it. The whole experiment consisted of 6 rounds, each round consisted of memorizing 2 sets of 15 words each. Words were randomly generated from a website online. I created imagery associated with each word and placed it in a locus.

One set was memorized by looking at the specific loci from 1 direction only, the other was viewed from 6 directions - front, left, back, right, up, and down.

The MPs used for this experiment were 360-degree views from google maps that were studied for 5 minutes just before memorizing the words. Rounds 1, 2, and 6 were made by first memorizing 15 words with SSP, then a 10-minute break and memorizing the other 15 words with MSP. The other 3 rounds were done in the opposite order.

Each set of words was checked to see how well it was remembered exactly 1 hour after first memorizing it and around 4 days afterward [+/- a couple of hours, however, that shouldn’t impact
the retention by much]. I counted any imagery that was remembered, even if the decoding ended up yielding a wrong result because this experiment focuses on the retention of the memorized imagery itself, however, decoding didn’t fail more than once or twice in some of the sets anyway. Rounds 1, 2, and 3 were done on the same day, round 4 was done on another day and rounds 5 and 6 were done on another day as well.

Results: Multiple Spatial Points [MSP] vs A Single Spatial Point [SSP]
Round 1)
Time: 03:21 min vs 02:15 min
Review in 1 hour: 13/15 vs 11/15
Review in 4 days: 14/15 vs 9/15

Round 2)
Time: 03:53 vs 03:17
Review in 1 hour: 11/15 vs 14/15
Review in 4 days: 14/15 vs 11/15

Round 3)
Time: 05:52 vs 03:45
Review in 1 hour: 13/15 vs 10/15
Review in 4 days: 15/15 vs 14/15

Round 4)
Time: 06:02 vs 04:05
Review in 1 hour: 11/15 vs 12/15
Review in 4 days: 13/15 vs 12/15

Round 5)
Time: 04:42 vs 03:14
Review in 1 hour: 14/15 vs 9/15
Review in 4 days: 14/15 vs 14/15

Round 6)
Time: 05:42 vs 03:08
Review in 1 hour: 13/15 vs 9/15
Review in 4 days: 13/15 vs 11/15

I would like to note that rounds 1 and 2 are a bit different than the rest for the following reason. When I was memorizing words on them when doing the SSP sets, I kept looking at the Google images and imagined events on them, however when I began doing MSP, I had to keep my eyes almost always closed so I could do it properly. Because of how “unstudied” the MPs used were, with only 5 minutes per MP, I can hypothesize that looking at an image while placing imagery there improved the retention of the SSP view. To balance that, I started keeping my eyes closed for all other sets in the next rounds [I did open my eyes between each locus to see it].

Interpreting the results

Time
By looking at the time for each of the rounds and sets, we can see that MSP usually ends up ~50% slower than an SSP. However, it is important to note that these are my personal results and I can say that I did have difficulties while using MSP. I believe that the time can be shortened by training. I also had to replay the full imagery from each spatial point. I did try to “speed it up” in my mind. I plan on having further tests with this in the future once I have trained it enough in my day-to-day use.

Review in 1 hour
Here, only round 2 appears to lean a lot toward SSP in terms of retention, round 4 also appears to be better for SSP but by a little bit. The other 4 rounds all go towards the MSP in terms of retention. That being said, it may be the case that the extra time spent on the MSP has allowed for that retention to happen, rather than the usage of MSP.

Review in 4 days
When reviewing in 4 days, the images from the MSP sets appeared to be more “clear” and more easily remembered than the SSP, but not by much. The retention [of MSP] appears to be better than that of SSP, except for round 5 where they are equal.

Retention of MPs
As a side effect of trying to learn MPs with MSP and then its imageries with MSP as well, I can remember the MPs themselves better. I have experienced something similar when trying to remember rooms and houses for my long-term projects. So it is possible that even if MSP is not used for remembering the actual images, it can be used for remembering MPs more easily.

Conclusion
Overall, it looks like MSP does potentially increase the retention of imagery. Unfortunately, some variables may have impacted the results, such as the additional time spent on the MSP [~50%]. I will conduct a new experiment in a month or two once I’ve spent more time using MSP in my day-to-day MP usage, where I can hopefully begin doing it faster and more easily. The amount of memorized words was also rather small [15 per set].

4 Likes

I do a version of this. I consider a room in a different state to be a different room. For instance, my kitchen with the table by one wall as opposed to the other wall, or my bedroom when it’s messy versus tidy.

1 Like

During the process of changing /adding viewpoints and possibly objects (the viewpoint inside the car will probably include small objects (the passenger seat, idk) that are related to the information one is trying to remember.
What does the car smell like inside, what does it feel like to be at such a speed, it makes you think it’s dangerous and you might crash, etc
It increases the connections.

1 Like

Very good question, perhaps you are referring to what’s called image streaming.

if I can point you toward something its Image Streaming on youtube by Mativier.

Essentially, when it comes to visualization the jar is never full, and you can increase the connections by diving deeper.

I do it all the time.

Images become loci themselves, for I stuff them with so much data, details and viewpoints it becomes its own thing.

So you are taking the individual places of a memory palace and tuning them into mini MP, increasing the amount of information there (I do that a lot). Eventually u might consider turn the places on the miniMPs into mini mini MP, and so on…like for ever inception
There are many things that certainly will go wrong, but I wont go into that like now(trust me, that’s a whole thing),I don’t like long posts.
Just go ahead and mess around with the idea, then came back here when u get questions.
Any ways… Yeah :+1: (⁠◠⁠ᴥ⁠◕⁠ʋ⁠)