Research
- M. T. Banich
III.
The modulation of attentional functioning by interaction between brain
regions
Although most researchers examining the neural basis of
attention conceive of it as a network of structures working in concert (e.g.,
Mesulam, 1981, Posner, 1992), the manner in which the interplay of these
various brain structures modulates attention remains unspecified.
Using behavioral methods as well as ERPs, I have found evidence for
the hypothesis that dynamic interactions between brain regions, in particular
interactions those between the cerebral hemispheres, have profound effects on
attentional functioning (see Banich, 1998 for a recent review).
This research is currently funded by a four-year NIMH grant.
This line of research has grown out of my comprehensive, systematic
and long-standing research program which examines how information is
integrated between the cerebral hemispheres.
One of the major mysteries facing cognitive neuroscience is how
information is integrated between different brain regions that are
specialized for different functions.
I have been examining this question by using interaction between the
cerebral hemispheres as a model system, since the hemispheres are
anatomically and functionally distinct.
This line of research has required me to invent a host of
investigatory paradigms (e.g., Banich and Belger, 1990; Passarotti and Banich,
submitted; Weissman and Banich, 1999) and to carefully consider
methodological issues (Banich and Shenker, 1992).
One of the major findings to come out of this research is that
interaction between the cerebral hemispheres can have properties that are
emergent, that is properties that cannot be surmised from the functioning of
each hemisphere in isolation (Banich and Karol, 1992).
This research makes it very clear that the whole is more than the sum
of the parts. My more recent
research has suggested that one of the important emergent properties of
interaction between the hemispheres is the modulation of attentional
functioning. Some of the main
findings of my research are:
1) The
processing capacity of the brain, which is limited and necessitates
attentional selection, can be enhanced by dividing information critical for
task performance between the hemispheres (as opposed to directing all the
critical information to one hemisphere).
We have found that when tasks are computationally complex, having both
hemispheres involved in processing, aids performance, an effect that is not
observed when the task is computationally simple (Banich and Belger, 1990;
Belger and Banich, 1992, 1998). Suggesting that this is a general attentional
phenomenon, the effect generalizes across response procedures (e.g., go/no-go
vs. choice RT), types of stimuli (e.g., alphanumeric characters vs. geometric
forms), particular types of decisions (e.g. arithmetic, categorical) (Banich,
Passarotti, and Chambers, 1994), spatial arrangement of the displays (Banich,
1985; Weissman, Banich, and Puente, undergoing revision), and modalities
(e.g., visual, tactile, auditory) (Passarotti and Banich, in preparation).
We have found that even when information is presented in central
vision (and hence is accessible to both hemispheres) and either within- or
between-hemisphere processing is possible, between-hemisphere processing is
invoked under computationally complex conditions, and within-hemisphere
processing is invoked under computationally simple conditions (Weissman and
Banich, 2000). The mechanism
that appears to allow for such an enhancement in performance is that by
dividing processing and then later on integrating the results, at least some
aspects of processing can be performed in parallel between the hemispheres (Banich
and Passarotti, submitted).
2) The
effect of interfering or distracting information can be reduced by
interaction between the cerebral hemispheres.
We have found a reduction in interference in a number of classical
paradigms of selective attention including the Navon
global-local paradigm (Weissman and Banich, 1999), the Stroop paradigm (Shenker
and Banich, submitted), and selective attention to one attribute of an item
(e.g., its form) while ignoring another (e.g., color) (Banich and Passarotti,
submitted) by dividing critical information across the hemispheres.
Our work has also helped to understand the mechanism by which such
effects occur. Although it has
been proposed that such a reduction in interference might occur because there
is selective of transfer between the hemispheres of only the task relevant
attribute (e.g., transfer of global but not local information) (Chiarello and
Maxfield, 1996) work in our laboratory suggests otherwise.
Our results indicate that as selective demands increase, there is
exchange between the hemispheres of both task relevant and irrelevant
attributes (Banich and Passarotti, submitted; Scalf and Banich, 1998),
suggesting that it is the increase in resources bought to bear on a task by
interaction between the hemispheres that aids in attentional selection.
3) Because
each cerebral hemisphere is a semi-autonomous system, central aspects of
processing can be facilitated when control or input is switched between the
hemispheres .
Many
theories of attention assume that there is a central bottleneck in processing
(e.g., Broadbent, 1958; Pashler & O’Brien, 1993).
We have found evidence that this bottleneck may not be totally
central, but that each hemisphere may have its own bottleneck.
We have investigated this issue by examining the attentional blink
phenomenon (Scalf, Banich, and Narechania, 1999), in which the processing of
one target item precludes processing of a second target item that is
presented shortly afterwards (e.g., within 150 msec).
The attentional blink can be reduced when the two critical stimuli are
directed to opposite hemispheres, suggesting that there is not a unified
central bottleneck.
4) The
differences in attentional control exhibited by each cerebral hemisphere
cause them to interact asymmetrically such that information given to the left
hemisphere affects processing of the right in a different manner than vice
versa.
One of
the clearest expressions of hemispheric asymmetry occurs with regards to
spatial attention. As clearly
demonstrated by hemi-neglect, the right hemisphere has an ability to direct
attention to both sides of space, whereas the left hemisphere has a strong
orienting bias to the opposite side of space (e.g., Weintraub and Mesulam,
1987; Kinsbourne, 1993). We have
found that these orienting biases affect the degree to which information
presented to one hemisphere can be insulated from that in the other.
Moreover, we have found that the degree of insulation is malleable, as
it is affected by the severity of the processing load.
It appears that the right hemisphere can insulate its processing from
distracting information presented to the left hemisphere under low load
conditions, whereas the left cannot. However,
under higher load conditions, distractors presented to the left hemisphere
have a more potent effect in disrupting processing of the right hemisphere
than vice versa (Passarotti and Banich, 1997).
These results contrast with Lavie’s (e.g. Lavie & Tsal, 1994)
theory of attentional selection which posits the opposite, namely that
distracting information will be more potent when overall attentional demands
are low. Our ERP studies provide evidence as to the stage of
processing at which such effects can occur.
We have found that information presented to the left hemisphere is
likely to affect the right hemisphere at earlier perceptual stages of
processing, whereas information from the right hemisphere is more likely to
affect the left hemisphere at response stages (Spencer, Banich and Coles,
1997; Spencer and Banich, 1998).