Mixing Memory

Gender Differences In Planning?

We all know that there are gender differences in neuroanatomy, as well as in some cognitive tasks (females tend to do better on memory and verbal tasks, men on spatial tasks) and both cognitive and emotional development, though it’s not clear how the cognitive/behavioral/developmental differences relate to the differences in neuroanatomy. Research on gender differences is often plagued by confounding variables such as sociological factors that are damn near impossible to control for. Jumping into the gender-differences game takes a lot of guts, or extreme naivete, then, because not only is your research going to be subject to increased scrutiny that comes with politically-charged rsearch, but it’s inevitably going to be inadequate as well. That’s certainly true of the latest entry into the gender differences literature, a paper by Boghi and 600 other people (neuroscientists travel in packs) in the November issue of NeuroImage on gender differences in planning in particular1. Before I get to the criticisms, though, I should probably describe the paper.

The study described in the paper uses the Tower of London task. In this task, there are three poles and a number of discs on the poles (usually 5 discs total). The goal of the game is to take the discs and move them so that they are in some pre-assigned arrangement, and to do so in as few moves as possible. The rules are that you can only move one disc at a time, and can only move a disc when there are no discs on top of it. Here’s a figure that shows an example of the setup, with the top image representing the goal, and the bottom image representing the starting configuration2:


This task has been used extensively to study planning, so it’s not surprising that Boghi at al. would choose it. In planning versions, people are asked to think about how to complete the task before they actually do it. Of course, they had participants strapped into fMRI machines, so it wasn’t really possible for the participants to actually move discs (or in their version, balls) on the poles. Instead, they had participants think about the minimum number of moves it would take to reach the goal state from the starting position, and then report this (in a whisper) to the experimenter. While they were figuring it out, pretty pictures were taken of their brains.

Now, both males and females do equally well on the Tower of London task generally, and Boghi et al. found no gender differences in task performance, but they weren’t interested in performance differences. They were interested in differences in brain pictures. When they contrasted activation during the performance of the task for males and females, they found greater activation in the female participants than the male participants in the dorsolateral prefrontal cortex (DPC; working memory) and right parietal cortex (attention). The male participants showed greater activation than the female participants in the precuneus, an area associated with spatial reasoning. Boghi et al. interpret these results as indicating that the female participants used more executive strategies for planning (working memory is considered an executive function), while the males relied more on spatial reasoning, an interpretation consistent with the gender differences I mentioned above. Boghi et al. argue that these differences reflect differences in planning strategies.

Buuuuuut there are problems. Putting aside the epistemological problems involved in making inferences about cognitive processes from neuroimaging data, the first problem with the study is that while the Tower of London task has been widely used to study planning, very little research has been conducted to determine exactly what cognitive functions are involved in performing the task. In fact, recent research has raised questions about whether planning is involved in the Tower of London task at all3. Furthermore, we do know that memory does plan an important role in the task4. It would not be unreasonable to interpret Boghi et al.’s results as simply showing that, because women tend to do better on tasks involving memory, when performing the Tower of London task, men more than women have to rely more on on-line spatial reasoning to perform as well. It might be interesting to learn that when mentally performing a task involving verbal, executive processes and visuo-spatial processes, women rely more on the verbal, executive ones, while men use more visuo-spatial ones, but I don’t think it is. I mean, we already knew that. So it’s not clear, then, that these results say anything about planning and gender, or anything new at all.

1 * Boghi A., Rasetti R., Avidano F., Manzone C., Orsi L., D’Agata F., Caroppo P., Bergui M., Rocca P., Pulvirenti L., Bradac G.B., Bogetto F., Mutani R., Mortara P. (2006). The effect of gender on planning: An fMRI study using the Tower of London task. NeuroImage, 33(3), 999-1010.
2 From Phillips, L.H., Wynn, V.E., McPherson, S. & Gilhooly, K.J. (2001). Mental planning and the Tower of London task. Quarterly Journal of Experimental Psychology, A, 54(2), 579-598.
4Phillips, L.H., Wynn, V., Gilhooly, K. J., Della Sala, S., & Logie, R. H. (1999). The role of memory in the Tower of London task. Memory, 7, 209-231.


  1. #1 Kevin McGrew
    November 22, 2006

    I just want you to know that I am a big fan of your blog and make frequent FYI posts re: your posts over at my blog — IQs Corner. Keep up the good work.

    Kevin McGrew

  2. #2 Free Operant
    November 22, 2006

    Two comments:

    First, for me, this study questions the value of fMRI imaging. The study shows we have different brain areas being activated while the same behavior is emitted. So, what have we learned?

    Second, I’m not a neuropsychologist, so I probably got this wrong. But, my impression is that “executive functions” is a pretty poorly defined concept. In fact, I vaguely recall some author excluding memory from that term. Instead that author used “executive functions” for higher level operations, such as judgment or insight.

  3. #3 Alvaro
    November 22, 2006

    Free Operant,

    2 very logical points! but let me offer some information to better understand the post.

    1) If every person showed different activations in “random” ways, yes, fMRI would be useless. But what I get from the post is that 1) men as a category displayed a pretty consistent (probably not statistically speaking, but I haven’t read the paper) pattern, 2) women as a category displayed a pretty consistent patter, 3) those 2 patterns were clearly different.

    2) Working memory is the ability to hold different pieces of information “online” in our minds, and is intimately associated with executive functions such as judgement and insight-you can not evaluate Pros and Cons and reach a decision without a “working” working memory. This is different from other types of memory, which, as you suggest, are not part of the executive functions. Some neuropsychologists are suggesting the use of “working attention” instead of “working memory” precisely to avoid this terminological confusion.

    Enjoy Thanksgiving

    NB1: Kevin, thanks for linking to this post!

    NB2: Mixing Memory, does the study say 1) what proportion of people found the solution, 2) what happened in their brain in the insight/ Aha! moment?

  4. #4 Chris
    November 22, 2006

    Free, “executive functions” is a pretty broad concept, and includes working memory, though not long-term memory. Women tend to do better on verbal working memory tasks, and the study I linked shows a connection between verbal working memory and the TOL task. Alvaro is correct in stating that working memory and long-term memory are probably very different things (in fact, different kinds of long-term memory are probably very different things), so it’s a bit confusing to use “memory” for both. I wrote the post quickly, and probably should have done a better job of indicating what I meant by “memory.”

    Alvaro, they do present the behavioral data in the form of a graph (it gives male, female, and total averages for each level of difficulty). Unfortunately, the paper is not available online without a subscription. If you don’t have access, I’d be happy to send you a copy in PDF format via email.

  5. #5 Chris
    November 22, 2006

    Oh, I forgot to add that the temporal resolution of fMRI probably isn’t good enough to know what happened at precisely the moment when they figured the solution out, and they didn’t really have participants indicate when they had figured out. The participants whispered the solution right before a task session ended.

  6. #6 Alvaro
    November 22, 2006

    Thanks Chris. If you can email me the paper, I’d appreciate. Happy Thanksgiving!

  7. #7 Mikko
    November 23, 2006

    Chris, this article and some of the comments above remind me of one thing I’m curious about. When papers report fMRI scan results, how consistent are the results typically? For example, in this case – how consistent is the difference between male and female images, or is there again some sort of distribution that heavily overlaps, but shows differences in the mean (as happens in many of the outward measures, such as IQ test scores).

    Perhaps, when you are talking about a paper, you could give us an indication of how strong the results are in this sense.

  8. #8 Chris
    November 23, 2006

    Hey Mikko, when writing these posts, I usually avoid giving some indication of variance, mostly because I’m afraid most people who read this won’t know what it means (not everyone takes stats). It’s even more difficult with imaging data, because the stats probably won’t make sense to anyone who doesn’t regularly read or do imaging stuff. If you’d like to see the data, though, I can send you a copy of the paper. The data for males and females are in Table 3.

  9. #9 John
    November 23, 2006

    I worry that Liz Spelke or Nancy Hopkins might be offended by the very premise of this study. You are taking a big risk by posting on this topic. I admire your courage.

  10. #10 Chris
    November 23, 2006

    John, that’s a bit unfair. Spelke, for example, sites tons of research on gender differences.

  11. #11 Mikko
    November 23, 2006

    Chris, perhaps you could make one post about the traditional way of looking at variances in cognitive sciences. As a non-cognitive scientist, I probably don’t need to go into detail in each of the papers you write about, but it would be very much beneficial to have some sort of baseline to compare to (or to know that there are big differences in how people use statistical data, if that’s the case).

  12. #12 Chris
    November 23, 2006

    Hmmm… variance is used in a lot of different ways, thought the most common are the parametric statistics (t-tests, ANOVAs, etc.). I suppose I could give a basic stats primer. Is that what you’re thinking of?

  13. #13 Rosemary
    November 25, 2006


    Would you email me the .pdf version that you mention to Alvaro?

New comments have been disabled.