What colour is an elephant
14 Oct 95
OUR knowledge of things in the world around us is
stored in networks of nerve cells all round our
brains. Researchers from the US National
Institute of Mental Health who have been looking
at where memories are stored now say that we
keep our knowledge about different properties of
objects, such as colour and movement, in
different places. And the knowledge about each
property is stored close to the part of the brain
that specialises in perceiving it.
People with brain damage occasionally suffer from
bizarre defects, such as being unable to
associate elephants with the colour grey while
having no difficulty perceiving and naming either
colours or elephants. Rather than having a single
file marked "elephant" in our mental filing
system, we seem to have a network of linked databases each
specialising in a particular kind of information. But when it comes to
identifying the brain areas responsible, brain damage can provide only
a rough guide: cases are rare, and the damaged areas are often large.
More detailed studies can be done by measuring blood flow in the
brain, using the scanning technique known as positron emission
tomography (PET). This allows researchers to identify peaks of
activity in different regions of normal people's brains while they carry
out particular mental tasks. Alex Martin and his colleagues at the
NIMH decided to use the technique to investigate knowledge about
colour and movement, because the areas of the brain that perceive
these attributes are already well known.
They showed people black and white line drawings of a number of
objects, and asked them to name either a colour or an action
associated with each one. For example, a picture of a pencil might
yield the responses "yellow" and "write".
Both tasks caused the brain to work harder in the prefrontal, parietal
and temporal lobes of the cortex, the outer layer of the brain that
deals with higher functions such as perception and language. Some
of this activity, however, was for subtasks such as word finding,
which come into play when naming both actions and colours. So
Martin and his colleagues computed the difference in brain activity
between the two tasks. Brain regions that responded equally to both
tasks yielded a difference of zero, while those that responded strongly
to one task and not the other showed a high value.
Their analysis revealed that the lower surfaces of each temporal lobe
contained an area that was most active when the subjects thought
about colour. This is close to a region other PET studies have
pinpointed as critical for colour perception. Similarly, an area near
the junction of the left temporal and occipital lobes was most active
when thinking up an action word - and this was immediately next to
the motion perception area (Science, vol 270, p 102).
"What's really important here is that knowledge is organised in a way
that's predictable from what we know about the organisation of
perception," says Martin. The results also support the idea that
perception of an object almost instantaneously makes available
everything we know about it. "We always see things in terms of the
meaning they convey," says Martin. "That's how we can identify them
so quickly."(see Diagram)
GEORGINA FERRY
From New Scientist magazine, vol 148 issue 1999, 14/10/1995, page
17
© Copyright New Scientist, RBI Limited 2001