A
very interesting question is whether there are particular parts of
the brain that are strongly associated with particular facets of
human mental functionality. Are there, for example, regions of the
brain that work much harder when you learn something, or remember
something, or feel something? The idea that there are such areas is a
hypothesis called localization.
Some
claim that this idea of localization is supported by brain imaging
studies. It is claimed that quite a few studies tell us about the
neural correlates of conscious experiences. In a typical study of
this type, people will have their brains scanned by some instrument
such as an MRI machine. Then scientists will look for certain parts
of the brain which showed more activity (such as blood flow) when
some particular type of mental activity was occurring.
But
there are reasons for thinking that such studies tell us very little.
For one thing, brain imaging studies on the neural correlates of
consciousness typically involve only small numbers of participants
(often fewer than 25). Making generalizations from such small
samples is dubious.
Also,
claims that particular regions of the brain show larger activity
during certain mental activities are typically not well-replicated in
followup studies. A book by a cognitive scientist states this (page
174-175):
The
empirical literature on brain correlates of emotion is wildly
inconsistent, with every part of the brain showing some activity
correlated with some aspect of emotional behavior. Those experiments
that do report a few limited areas are usually in conflict with each
other....There is little consensus about what is the actual role of a
particular region. It is likely that the entire brain operates in a
coordinated fashion, complexly interconnected, so that much of the
research on individual components is misleading and inconclusive.
There
have been statistical critiques of brain imaging studies. One
critique found a common statistical error that “inflates
correlations.” The paper stated, “The underlying problems
described here appear to be common in fMRI research of many kinds—not
just in studies of emotion, personality, and social cognition.”
Another
critique of neuroimaging found a “double dipping” statistical
error that was very common. New
Scientist reported a software problem, saying “Thousands of fMRI
brain studies in doubt due to software flaws.”
Considering
the question of “How Much of the Neuroimaging Literature Should We
Discard?” a PhD and lab director states, “Personally I’d say I
don’t really believe about 95% of what gets published...I think
claims of 'selective' activation are almost without exception
completely baseless ”
Another
huge reason for being skeptical about brain imaging studies is that
such studies very often use very misleading visual presentations,
creating false impressions. What very frequently goes on is
something like this. A series of brain scans will show a very small
difference between brain activity in different parts of the brain – typically only 1%. A Stanford
scientific paper on fMRI uses this 1% figure, making an exception
only for small parts of the brain (“visual and auditory cortices”)
associated with seeing and hearing. The paper makes this
generalization: “While cognitive effects give signal changes
on the order of 1% (and larger in the visual and auditory cortices),
signal variations of over 10% may arise from motion and other
artifacts in the data.” In other words, people moving their heads
(and other misleading signals) may create the impression that there
is a higher variation in the non-sensory parts of the brain, but the
real variation in the signal changes is only something like 1%. A
similar generalization is made in this scientific discussion, where
we are told the following:
For example, most cognitive experiments should show maximal contrasts of about 1% (except in visual cortex), hence, if estimates for a single subject are much larger than that, then the estimates are likely to be bad. Poor estimates can arise from head motion, or sporadic breathing patterns by the subject, or sometimes from a poor design matrix that is ill-conditioned.
For example, most cognitive experiments should show maximal contrasts of about 1% (except in visual cortex), hence, if estimates for a single subject are much larger than that, then the estimates are likely to be bad. Poor estimates can arise from head motion, or sporadic breathing patterns by the subject, or sometimes from a poor design matrix that is ill-conditioned.
But
again and again studies on neural correlations will produce a visual
which grossly exaggerates this very small difference of 1%, making it
look like a great big difference. This is lying with colors.
To
explain why this is highly deceptive, let's consider some examples of
displaying visual information: an honest presentation and a
misleading presentation. Imagine you have a home for sale, and you
are preparing a web page or brochure that describes your house's selling
points. One of the key factors in selling houses is the quality of
the local school district. Anyone with children will want to buy a
house in a neighborhood with better schools.
Imagine
you've got the average school stores for your home's school district,
and the data looks like this:
District | Average Reading Score |
District 1 | 80 |
District 2 | 80 |
District 3 | 80 |
District 4 (yours) | 81 |
District 5 | 80 |
District 6 | 80 |
District 7 | 80 |
District 8 | 80 |
Now
imagine you wanted to present a school district map highlighting the
higher score of your home's school district. For you to honestly
present such information, you would have to follow this rule: the
difference in the color shade should be no different than the
differences in the data that you have.
So
if you were to present an honest school district map, color-coded
school scores, it would have to look something like this:
This
would be an honest map. It honestly indicates only a very slight
difference between the school scores in your home's school district
and the nearby districts. But you might be tempted to present the
data differently. You might present it like the map below.
This
map would be better from the standpoint of selling your house, since
it would leave someone with the impression that your house's school
district has much higher scores than the nearby districts. But given
the actual data showing only a 1% difference, it would be utterly
misleading to present a map like this. The map would incorrectly give
someone the idea that your home's school district had scores maybe
30% better than surrounding districts. Presenting a map like this
would be an example of lying
with colors.
It
is exactly such lying with colors that goes on again and again in
brain imaging studies on the neural correlates of consciousness.
Again and again, such studies will show visuals that depict
differences of only 1% or less between blood flow in different
regions of the brain. But such regions will be shown as red regions
in brain images, with all of the other areas having a grayish “black
and white” color. When you see such an image, you inevitably get the
impression that the highlighted part of the brain has much higher
activity than other regions. But such a conclusion is not what the
data is showing.
So,
for example, a study finding 1% higher brain activity in a region
near the corpus callosum (under some activity that we may call
Activity X) might release a very misleading image looking like the
image below, in which the area of 1% greater activity is colored in
red.
But
such an image is lying with colors. If there is only a 1% greater
activity in this region, an honest diagram would look like the one
below.
With
this diagram, the same region shown in red in the first diagram is shown as only 1% darker. You
can't actually tell by looking at the diagram which region has the 1%
greater activity when Activity X occurs. But that's no problem. The
diagram above leaves the reader with the correct story: none of the
brain regions differ in activity by more than 1% when Activity X
occurs. Contrast this with the first image, which creates the very
misleading idea that one part of the brain is much more active than
the others when Activity X occurs.
You
might complain that with such a visual, you cannot tell which regions
have the slightly greater activity. But there are various ways to
highlight particular regions of a brain visual, such as circling,
pointing arrows, outlining, and so forth. For example, the following
shows a region of high activity without misleading the viewer by
creating the impression of much higher activity:
The misleading diagrams of brain imaging studies seem all the more appalling when you consider that the images in such studies are typically the only thing that laymen use to form an opinion about localization in the brain. The text of brain imaging studies is typically written in thick jargon that only a neuroscientist can understand. Frustrated by this very hard-to-understand jargon and unclear writing, every layman reading these studies forms his opinions based on the visuals. When such visuals deceive us by lying with colors (as they so often do), the whole study ends up being something that has the effect of creating misleading ideas.
At
4:47 in this online course, we are told that the blood changes in the
brain are quite small when observed with an MRI, between 0.1% and 5%
(based on two previous comments I quoted, the higher value seems to
be only found in the auditory cortex or the visual cortex). The
speaker in this course tells us that if we were to just watch a movie
of the fMRI scans, we wouldn't be able to notice the changes between different brain regions.
But the visuals in neural correlation studies misrepresent this
minimal change, making it look like a great big change. These cases
of lying with colors give readers a very misleading impression that
brain activity involving thinking and memory recall is very
localized, and tightly correlated with mental activity.
How
closely your brain is correlated with your mind is important
from a philosophical standpoint. If different parts of our brains
surge dramatically with blood when we think or recall memories, that
is a point favoring the idea that your mind is just a product of your
brain. But if different parts of your brain look pretty much the same
when you think or recall memories, that's a point in favor of the
idea that your mind and memories may involve something much more than
your brain, perhaps some soul or some higher cosmic consciousness
infrastructure. The actual data from brain imaging is the second of
these cases: different parts of the brain show about the same activity when thinking or memory
recall occurs. But by doing neural correlation studies that visually
make tiny changes look like huge changes, our neuroscientists almost
seem to be trying to fool us into thinking that a very different
thing is going on, that particular parts of our brain light up dramatically when
our higher mental functions are engaged. To such neuroscientists we
should say: visually represent your own data honestly, and stop lying
with colors.
No comments:
Post a Comment