Object substitution masking (OSM) is a recently re-
ported form of visual masking that does not depend
on the factors established as important in conventional
masking (e.g., the degree of spatial overlap or abutting
contour). In OSM, a target can be rendered impercep-
tible by a mask consisting of just four dots surrounding
the target, or, in some circumstances, a single dot pre-
sented at a distance from the target location. For OSM
to occur, at least two conditions do need to be fulfilled:
that the target not be the focus of attention, and that the
mask remain present after target offset (Di Lollo, Enns,
& Rensink, 2000).
Initial accounts of OSM argued that it arose as a con-
sequence of iterative reentrant exchanges in the cortex
that occur as part of the normal processes of the visual
system forming a stable representation based on current
input. The presence of the trailing mask surrounding
the target, it was argued, dramatically reduced the pos-
sibility that the stable representation that arises from the
iterative exchanges would include the target as well as
the mask (Di Lollo et al., 2000). In the original substitu-
tion account of OSM, the assumption was that masking
consisted of competition for consciousness between two
separate object representations. The more recent object
updating account of OSM (Enns, Lleras, & Moore, 2009;
Lleras & Moore, 2003; Moore & Lleras, 2005) proposes
that this competition occurs within a single object repre-
sentation. This account focuses on the interactions that
occur at the object token level of representation, hypoth-
esized midlevel representations considered responsible
for maintaining object integrity by linking together the
different object attributes over changes in time and space
(Kanwisher & Driver, 1992). In the object updating ac-
count, OSM is viewed as a consequence of the visual
system failing to individuate target and mask as distinct
perceptual objects, meaning that the attributes of the tar-
get and of the mask are initially bound within the same
object token representation. If, under such circumstances,
the mask lingers in the display after target offset, the rep-
resentation will be updated, so the joint target and mask
features are overwritten by those of the mask alone, ren-
dering the target features imperceptible.
Key evidence for the object updating account of OSM
was presented by Lleras and Moore (2003). In a critical ex-
periment, a target search array was briefly presented around
a fixation point. The array was replaced after a variable
interstimulus interval (ISI) by a frame consisting of single
dots (the masks) presented at locations eccentric from tar-
get array items (see the upper portion of Figure 1 for il-
lustration): With a short (17- to 34-msec) ISI, the onset of
the mask frame generated a motion signal connecting the
dots with the corresponding items in the search array; with
the long (216- to 233-msec) ISI, this connecting apparent
motion (AM) was not observed. Importantly, masking was
found in the short, but not in the long, ISI condition, when
measured against a no-mask control condition. The con-
necting AM was interpreted as being critical for masking
to occur under these conditions. The presence of the AM
suggested that, at the token level, the visual system was
representing the target and mask items as a single object.
The masking associated with this could accordingly be
viewed as caused by the mask overwriting the attributes
of the target associated with the same object token, since it
was seen to transform over time and space. The absence of
the connecting AM with a long ISI conversely suggested
that target and mask were represented as distinct objects
737 © 2010 The Psychonomic Society, Inc.
Object substitution masking and the
object updating hypothesis
MICHAEL PILLING AND ANGUS GELLATLY
The Open University, Milton Keynes, England
The object updating hypothesis of object substitution masking proposes that the phenomenon arises when the
visual system fails to individuate target and mask at the level of object token representations. This hypothesis
is tested in two experiments using modifications of the dot mask paradigm developed by Lleras and Moore
(2003). Target–mask individuation is manipulated by the presentation of additional display items that influence
the linking apparent motion seen between a target and a spatially separated mask (Experiment 1), and by the
use of placeholders that maintain the target object’s presence during mask presentation (Experiment 2). Results
in both cases are consistent with the updating hypothesis in showing significantly reduced masking when the
conditions promoted target–mask individuation. However, in both experiments, some masking was still present
under conditions of individuation, an effect we attribute to attentional capture by the mask.
Psychonomic Bulletin & Review
2010, 17 (5), 737-742
doi:10.3758/PBR.17.5.737
M. Pilling, mpilling@brookes.ac.uk