s
sh
car
tree
ber
e o
n P
Update
TICS-896; No. of Pages 2individual RFs but of populations. Saccade targeting is
more accurate than the size of individual fields in saccade
control areas [4] and more similar to the size of attentional
selection areas. We did depict the models of remapping
(e.g. [5]) as relying on individual cells but that was an
oversimplification. Clearly, saccade and attention targets
must be indexed by a profile of activity across many
responding units and this entire profile would be subject
to remapping. Even if a stimulus would remain within the
RF of a single cell after a saccade, the population of
responding cells would shift.
Second, Mayo and Sommer ask if there is activity that is
specific to the attention pointer either in single neurons,
local microcircuits or more global networks. We assumed
that the function of the attention pointer was inherent in
the well-documented properties of saccade control areas
that specify the target location and provide, through down-
ward projections, attentional benefits at the corresponding
locations (review in [6]), instantiating the essential func-
tions of spatial attention. In this view, the attentional
benefits are an obligatory consequence of activity in the
saccade areas, not an optional related activity. When a
saccade is programmed, the activity peaks for current
targets are remapped to compensate for their upcoming
shift in retinal locations. As the activity peaks of attended
targets shift to new locations, their attentional benefits
shift with them. Here ‘attended’ means nothing more or
less than the object or features at the retinal location
corresponding to an activity peak. We argued, in particu-
lar, that remapping is not a RF shift (Box 1) but an
anticipatory response. This does raise the question of
how the downward projections arise and how accurately
they target neurons in earlier cortices. Nevertheless, our
labeling of this localized performance advantage as an
‘attentional pointer’ only specifies that the activity ident-
ifies the location of the target to which performance
benefits are provided.
Finally, along with Mayo and Sommer, we feel it is
crucial to determine to what extent the individual neurons
in saccade centers have featural specificity independently
of task relevance and to what extent the performance
benefits from downward projections can be tuned to target
features. This is an ongoing exploration and although we
claimed it would be more plausible that remapping of
activity did not convey feature information, the final
answer will emerge from ongoing work in our lab and
elsewhere.
References
1 Mayo, J.P. and Sommer, M. (2010) Shifting attention to neurons.
Trends Cogn Sci., doi:10.1016/j.tics.2010.06.003
2 Cavanagh, P. et al. (2010) Visual stability based on remapping of
attention pointers. Trends Cogn Sci 14, 147–153
3 Intriligator, J. and Cavanagh, P. (2001) The spatial resolution of visual
attention. Cogn Psychol 43, 171–216
4 Carpenter, R.H.S. (1988) Movements of the Eyes, Pion
5 Keith, G.P. and Crawford, J.D. (2008) Saccade-related remapping of
target representations between topographic maps: a neural network
study. J Comput Neurosci 24, 157–178
6 Awh, E. et al. (2006) Visual and oculomotor selection: links, causes and
implications for spatial attention. Trends Cogn Sci 10, 124–130
7 Anton-Erxleben, K. et al. (2009) Attention reshapes center-surround
receptive field structure in macaque cortical area MT. Cereb Cortex 19,
2466–2478Corresponding author: Cavanagh, P. (patrick.cavanagh@parisdescartes.fr).
that saccades easily do. Moreover, changes in RFs in middle
temporal area (MT) and V4 occur around the focus of attention [7],
as well as around a saccade target [9]; the important point is that eye
movements are not necessary for these RF modulations to be
observed. By contrast, remapping is only seen when an eye
movement is executed, and does not occur for attention shifts
[10], indicating it serves a separate function. The changes in MT and
V4 have been linked to enhancement of processing at attended
locations; remapping maintains that processing enhancement at
spatially appropriate locations by shifting activity peaks to the new
retinotopic location of attended locations when the eyes move.
1Letters Response
Attention Pointers: Respon
Patrick Cavanagh1,2, Amelia R. Hunt3, Ara
1 Laboratoire Psychologie de la Perception, Universite´ Paris Des
2Vision Sciences Laboratory, Harvard University, 33 Kirkland S
3School of Psychology, William Guild Building, University of A
4McGovern Institute for Brain Research, Massachusetts Institut
5Department of Psychology, New York University, 6 Washingto
Mayo and Sommer [1] raise several interesting questions
about our opinion article [2]. First, they propose that the
receptive field (RF) sizes in saccade control areas are too
large to support the localization of attentional benefits
seen in behavioral studies. Clearly the RF sizes in these
areas are large, but attentional resolution is correspond-
ingly extremely crude [3]. However, the RF sizes in frontal
eye fields (FEF) and lateral intraparietal (LIP) area are
perhaps twice as large as the corresponding ‘attentional
field’ at the same eccentricity. On this point we agree with
Mayo and Sommer that localization cannot be a function ofe to Mayo and Sommer
Afraz4 and Martin Rolfs5
tes, 45 rue des Saints Pe`res, 75006 Paris, France
t, Cambridge, MA 02138, USA
deen, Aberdeen, AB24 2UB, UK
f Technology, 43 Vassar Street, Cambridge, MA 02139, USA
lace, New York, NY 10003, USA
Box 1. Attention-driven RF shifts vs. remapping
Mayo and Sommer point out that the centers of large RFs at
intermediate levels do shift toward an attentional focus (e.g. [7]).
Saccade targets clearly set up an attentional focus, so Mayo and
Sommer’s point naturally raises the possibility that the remapping
shifts are related to these attention-driven shifts (see also [8]).
However, we argue that these attention-driven effects cannot
generate RF shifts of the magnitude required for remapping; they
are much too small. A RF must necessarily shift 100% of the saccade
magnitude to match the remapping results whereas attention-driven
RF shifts are at best 30% on average of the separation from target to
attentional focus [7] and never reach the 50 to 90 deg displacements

8 Zirnsak,M. etal. (2010)Thespatial distributionof receptivefield changes
in a model of peri-saccadic perception: Predictive remapping and shifts
towards the saccade target. Vision Res. 50, 1328–1337
9 Tolias, A.S. et al. (2001) Eye movements modulate visual receptive
fields of V4 neurons. Neuron. 29, 757–767
10 Duhamel, J.R. et al. (1992) The updating of the representation of visual
space in parietal cortex by intended eyemovements.Science 255, 90–93
1364-6613/$ – see front matter  2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tics.2010.06.004 Trends in Cognitive Sciences xx (2010) 1–2
Update Trends in Cognitive Sciences Vol.xxx No.x
TICS-896; No. of Pages 22