2013. Vol.4, No.11, 823-826
Published Online November 2013 in SciRes (
Open Access 823
Illusory Upward Self-Motion Results in a Decrease in Perceived
Room Temperature
Takeharu Seno1,2,3, George H. Van Doorn4
1Institute for Advanced Study, Kyushu University, Fukuoka, Japan
2Faculty of Design, Kyushu University, Fukuoka, Japan
3Research Center for Applied Perceptual Science, Kyushu University, Fukuoka, Japan
4School of Applied Media and Social Sciences, Monash University, Melbourne, Australia
Received August 21st, 2013; revised September 22nd, 2013; accepted October 23rd, 2013
Copyright © 2013 Takeharu Seno, George H. Van Doorn. This is an open access article distributed under the
Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any
medium, provided the original work is properly cited.
Purpose: Stationary observers often experience illusory self-motion (vection) when they are exposed to
large patterns of optic flow. The effect of different temperatures on the strength of vection was investi-
gated. Method: Eleven participants were exposed to visual stimuli that induced illusory motion (up, down)
in three room temperatures (26˚C - 27˚C, 21˚C - 22˚C, 5˚C - 6˚C). Participants rated (a) the vection mag-
nitude, and (b) the room temperature (twice; before and after vection). Results: Upward vection was rated
as stronger than downward vection in the 26˚C - 27˚C temperature. In addition, after experiencing upward
and downward vection, subjective ratings of room temperature decreased and increased, respectively,
when the room temperature was 26˚C - 27˚C. This effect was not observed when the room was 5˚C - 6˚C.
Conclusion: These results suggest that a cross modal association exists between the direction “up” and
26˚C - 27˚C temperatures.
Keywords: Vection; Temperature; Vision; Illusory Self-Motion
Exposure to a visual motion field that simulates the retinal
optical flow generated by self-movement commonly causes the
perception of the subjective movement of one’s own body. This
phenomenon is known as “vection” (Fischer & Kornmuller, 1930).
For example, when a stationary person observes a train begin-
ning to move, they are likely to perceive that they are moving
in the opposite direction to the motion of the train. This pheno-
menon is known as the “train illusion”, and provides a good
example of vection (e.g. Seno & Fukuda, 2012). Some recent
studies reported that vection can modulate aspects of perception
and cognition, e.g. daydreaming (Miles, Karpinska, Lumsden,
& Macra, 2010), attention (Seno, Ito, & Sunaga, 2011a), time
perception (Seno, Ito, & Sunaga, 2011b), perception of num-
bers (Seno, Taya, Ito, & Sunaga, 2011), and visual illusions
(Fukuda & Seno, 2011; Fukuda & Seno, 2012). Vection is also
affected by personality traits such as narcissism (Seno, Yamada,
& Ihaya, 2011). In this study, we assessed for the first time
whether vection can be modulated by temperature, or can have
a modulating influence on perceived temperature.
Soto-Faraco, Spence and Kingstone (2004) suggested that a
feature in one sensory modality can be associated with a feature
in another sensory modality, i.e. a cross-modal correspondence.
Although different sensory organs receive different inputs, in-
teractions between sensory modalities occur after these inputs
have been recoded at post-perceptual levels (Marks, 2004); re-
coding activates a representation that captures elements of each
input that are common across modalities (Martino & Marks,
2000). Consequently, reaction times (RTs) to a simultaneously
presented high-pitched tone and a light coming from above
might be shorter, for example, than RTs to a low-pitched tone
and a light coming from above; the recoded inputs of the for-
mer pair share the post-perceptual format “high”, whereas the
later do not.
Similar to Mark’s (2004) work, Seno, Ito, Sunaga, Hasuo,
Nakajima and Ogawa (2011) recently proposed a consistency
hypothesis that predicts that visually-induced self-motion will
be enhanced when inputs from one (or more) of the non-visual
senses are consistent with visual simulation. Consistent with the
predictions of this hypothesis, somatosensory stimulation gen-
erated by adding air-flow to a stationary observer’s face sig-
nificantly (p < .05) enhanced visually-simulated forward mo-
tion (Seno, Ogawa, Ito, & Sunaga, 2011). In addition, vection
has been facilitated by vibrations on the body and auditory cues
that are consistent with visual rotation (Riecke, Schulte-Pelkum,
Caniard, & Bülthoff, 2005; Riecke, Feuereissen, & Rieser, 2008).
Crossmodal correspondence, or consistency, seems to be non-
arbitrary and accumulated through repeated exposure to pairs of
stimuli (Williams & Bargh, 2008). As we go through life expe-
riencing wind against our face, for example, it becomes paired
with forward motion.
This idea can be applied to temperatures and spatial direc-
tions (up/down) which may be semantically associated in eve-
ryday language (e.g. “hot air rises”). If a correspondence exists
between “hot” temperatures and “upward” direction (and vice
versa for “cold” and “down”) then vection may be influenced
by temperature; specifically, upward vection would be stronger
when the temperature of a room is hotter than usual, and vice
versa for colder temperature. In addition, illusory change in
room temperature may arise as a result of perceiving vection.
Thus, we tested the following hypotheses:
H1: In a 26˚C - 27˚C room, upward vection will be stronger
than illusory self-motion in other directions.
H2: Subjective ratings of room temperature will be hotter af-
ter experiencing upward vection, and cooler after downward
Eleven volunteers (seven females) took part in the experi-
ment. The participants had a mean age of 24.8 years (SD = 3.7
years). All participants had normal or corrected-to-normal vi-
sion and none of them reported visual or vestibular abnormali-
ties. The experiments were pre-approved by the ethics commit-
tee of Kyushu University, and written informed consent was
obtained from each subject prior to participating.
Materials and Stimuli
The up/down vection stimulus was a white vertical sinusoid-
dal grating whose luminance was horizontally modulated. Mo-
tion displays subtended a visual area of 72˚ (horizontal) × 57˚
(vertical) when viewed from 570 mm in front of a television
screen. Upward/downward vection was induced by moving the
grating (spatial frequency: .1 cycle/deg; mean luminance: 18
cd/m2; Michelson contrast: 80%) down/up, respectively, at a
speed of ~20 deg/sec. We used virtually continuous motion in
the minimum motion technique that was essentially the same as
that used in Cavanagh, MacLeod and Anstis (1987). Each up/
down motion stimulus consisted of 13 images presented se-
quentially and repeatedly. The gratings moved in only one di-
rection for the duration of each trial, which was fixed at 30 sec.
There was also a static (i.e. control) grating condition.
The experiment was conducted in a 7 m × 7 m × 2.6 m dark-
ened room. There were three room temperature conditions, i.e.
normal (21˚C - 22˚C), hot (26˚C - 27˚C), and cold (5˚C - 6˚C).
Room temperature was modulated and kept constant by the two
air conditioners. Room temperature was certificated by the ther-
Prior to the presentation of stimuli, participants were asked to
sit down; they remained seated for the entire experiment. Be-
fore each stimulus combination was presented (e.g. 5˚C - 6˚C
temperature/upward grating pattern) participants were asked to
estimate the room temperature to one decimal point (e.g. 23.5˚C).
They were also instructed to press a button during each trial if/
when they experienced vection, and were to keep the button de-
pressed for as long as the experience lasted. If vection ceased,
or became ambiguous, they were to release the button. Duration
and latency of vection were recorded as dependent variables.
Latency was defined as the time interval between the onset of
the visual stimulus and the time at which the participant pressed
the button. Duration was calculated as the total time that the
button was pressed until it was released.
Following the instructions a motion stimulus was presented.
After the 30 sec stimulus presentation period ended, and after
participants released the button to signify that vection had ceas-
ed, participants rated the strength of vection by verbally stating
a number from 0 (no vection) to 100 (very strong vection); the
experimenter recorded this number on a piece of paper. Par-
ticipants then estimated the room temperature again, and the dif-
ference between pre- and post-trial estimations was calculated.
There were two motion directions and three room tempera-
ture conditions. There were four trials per condition, and thus a
total of 24 trials per participant. The presentation order of trials
was randomized. Each temperature condition lasted approxi-
mately 20 mins. Thus the experiment had a total duration of ap-
proximately 60 mins for each participant. All conditions were
interior-group design.
It is known that vection can be modulated by an experimen-
ter’s instructions or demands (Palmisano & Chan, 2004). Thus,
we carefully instructed the participants regarding their task with-
out giving them any suggestion which may lead to a cognitive
bias about the consistency hypothesis. Furthermore, several con-
trol conditions were included in an attempt to negate potential
extraneous variables.
A 3 (temperature: 21˚C - 22˚C, 26˚C - 27˚C, 5˚C - 6˚C) × 3
(grating motion: upward, downward, static) repeated-measures
ANOVA on subjective ratings of vection strength revealed a
significant main effect of temperature [F(2,20) = 21.62, p
< .0001, η2 = .16]. Vection was weaker in the 26˚C - 27˚C con-
dition than in the 21˚C - 22˚C and 5˚C - 6˚C cold conditions
(see Figure 1). This is supported by the duration and latency
data which were significantly shorter and longer, respectively,
in the 26˚C - 27˚C condition relative to the 21˚C - 22˚C and 5˚C
- 6˚C conditions [duration: F(2,20) = 6.83, p = .005, η2 = .06;
latency: F(2,20) = 13.62, p = .0002, η2 = .21].
The main effect of the grating motion was significant [mag-
nitude: F(2,20) = 66.62, p < .0001, η2 = .16; duration: F(2,20) =
86.29, p = .005, η2 = .70; latency: F(2,20) = 54.36, p < .0001, η2
= .55]. Bonferroni-corrected post-hoc tests revealed significant
differences between the static grating and the others (p < .05).
Importantly, the direction of illusory self-motion experienced
by each participant was consistent with previous research (e.g.
Seno, Ito, & Sunaga, 2009).
The interaction between temperature and grating motion was
significant [magnitude: F(4,40) = 9.36, p < .05, η2 = .10; dura-
tion: F(4,40) = 3.79, p < .05, η2 = .03; latency: F(4,40) = 12.19,
p = .002, η2 = .16]. Bonferroni-correct post-hoc comparisons
revealed that upward vection was stronger than downward vec-
tion in the 26˚C - 27˚C condition (p < .05), and thus H1 was
A second 3 (temperature) × 3 (grating motion) repeated-mea-
sures ANOVA on the difference between perceived tempera-
tures pre- and post-stimulus presentation revealed no significant
main effect of temperature [F(2,20) = 1.62, p = .22; η2 = .03]
(see Figure 2). The main effect of grating motion was signifi-
cant [F(2,20) = 7.17, p = .0045, η2 = .08]. Post-hoc tests
(Holm’s Sequentially Rejective Bonferroni Procedure) revealed
significant differences between down and the other two condi-
tions [up vs. down, t(10) = 2.98, p = .04; Cohen’s d = .73; down
vs. static, t(10) = 2.48, p = .03; d = .31]. There was also a trend
towards significance between the up and static conditions, t(10)
= 2.05, p = .06; d = .29. The interaction was significant [F(4,40)
= 2.64, p < .05, η2 = .12]. Bonferroni-corrected post-hoc com-
Open Access
Figure 1.
Strength of vection. The unit of measure-
ment for latency and duration were seconds,
while magnitude was a rating from 0 to 100.
There was no vection in any static condition,
and hence the “spaces” in the figures.
Figure 2.
The perceived change in temperature after vection.
parisons revealed that perceived room temperature was modu-
lated by motion stimuli only in the 26˚C - 27˚C condition (p
< .05).
Vection refers to the experience of illusory self-motion in re-
sponse to large patterns of optic flow. The main issues explored
in this study were whether (a) vection is influenced by changes
in room temperature, and (b) perceived room temperature is
changed by experiencing vection. Remarkably, the interaction
between temperature and grating motion was significant, and
thus vection seems to be influenced by room temperature. Fur-
ther, and contrary to H2, upward vection produced a perceived
drop in room temperature and downward vection resulted in a
perceived rise, but only in the 26˚C - 27˚C condition.
We speculate that, as upward vection was stronger than
downward vection in the 26˚C - 27˚C condition, there is a cor-
respondence between heat and upward motion that exists in clo-
sed spaces, e.g. rooms and buildings where hot air rises. How-
ever, in the 5˚C - 6˚C condition, downward vection was not sig-
nificantly stronger than upward vection. It is somewhat surpris-
ing that an association exists only between 26 - 27˚C and “up”,
and not between 5˚C - 6˚C and “down”. Although speculative,
this unidirectional finding may be explained by the work of
Riecke et al. (2008) who showed that the possibility of actual
self-motion facilitates vection. In their experiment auditory cir-
cular vection was enhanced by suspending participants above
the ground. In our experiment, participants’ feet touched solid
ground and they, presumably, knew that actual “downward”
motion was impossible, whereas given the free space above each
participant’s head it was (theoretically) possible for them to
move up into this free space.
To examine this hypothesis, we conducted an informal ob-
servation with 4 naïve volunteers in which their feet were kept
off the ground, and thus upward and downward motion was
(theoretically) possible; we, again, measured strength of vection.
The results were similar to those of the main experiment, i.e.
upward vection was stronger than downward vection in the
26˚C - 27˚C condition. Thus, the idea that contact between
one’s feet and solid ground inhibits vection is not supported by
our informal observations. That said, even though the partici-
pants’ feet did not touch the ground, their buttocks did touch
the chair. Further experiments are planned to determine whe-
ther manipulating the procedure shows correspondence effects
in the other conditions. The current results suggest that illusory
self-motion is stronger when there is a correspondence between
two stimuli and one knows that physical motion is possible.
Contrary to H2, upward vection produced a perceived drop in
room temperature while downward vection resulted in a per-
ceived rise, but only in the 26˚C - 27˚C condition. It is possible
that participants were aware of the fact that hot air rises and,
after experiencing illusory self-motion in the 26˚C - 27˚C tem-
perature condition, they expected an increase in room tempera-
ture. As the actual temperature remained constant after experi-
encing vection, because no physical positional change took place,
there was a discrepancy between expected and actual room
temperatures, which resulted in the room being judged colder
than it actually was. However, given downward vection was
weaker than upward vection in the 26˚C - 27˚C and, here, down-
ward vection resulted in a perceived increase in room tempera-
ture, this explanation is speculative. It might be that the hot
temperature (26˚C - 27˚C) and vection stimuli used here contain
some unknown features that correspond, whereas the cold tem-
perature (5˚C - 6˚C) and vection stimuli did not. There is some
evidence (see Pilcher, Nadler, & Busch, 2002) that heat can be
more attention-getting than can cold and, as such, it may be that
a hotter room shifts attention to temperature, and thus partici-
pants became sensitive to it.
It is known that exposure to visual stimuli like those used
here can create nausea in participants, and common autonomic
reactions are cold sweating and reduced peripheral blood flow.
These reactions are especially common among healthy, young
people who do not suffer from vestibular disorders. These are
natural autonomic responses and a shift in the autonomic nerv-
ous system towards a sympathetic reaction. Although each ex-
Open Access 825
Open Access
perimental trial lasted only a short time, there were repeated
exposures and the autonomic response is quick, especially if
trials are repeated with short inter-trial intervals. That said, for
autonomic responses to have been a confounding variable the
reported influence of temperature on vection would have to be
explainable having found that autonomic responses occurred in
the 26˚C - 27˚C temperature/downward motion (upward vection)
condition but not in the 26˚C - 27˚C temperature/upward mo-
tion (downward vection) condition, and that this increased the
perceived strength of vection in the former condition. There is
no evidence to support such a view.
In conclusion, we have demonstrated the existence of a new
crossmodal correspondence, namely an association between tem-
perature and directional up; vection modulated perceived tem-
perature, and was modulated by actual temperature. As Parise
and Spence (2012) argue, this correspondence might reflect the
natural correlation between physical properties of the world, i.e.
warmer air moves upwards relative to cooler air.
This work is supported by the Program to Disseminate Ten-
ure Tracking System, MEXT, Japan.
Cavanagh, P., MacLeod, D. I., & Anstis, S. M. (1987). Equiluminance:
Spatial and temporal factors and the contribution of blue-sensitive
cones. The Journal of the Optical Society of America, 4, 1428-1438.
Fischer, M. H., & Kornmuller, A. E. (1930). Optokinetic ausgeloste Be-
wegungs-wahrnehmungen und optokinetinetisher Nystagmus. Jour-
nal of Comparative Neurology, 41, 273-308.
Fukuda, H., & Seno, T. (2011). Shrinking neighbors: A quantitative
examination of the ‘Shrinking Building’ illusion. Seeing Perceiving,
24, 541-544.
Fukuda, H., & Seno, T. (2012). Healin’ groovy: Movement affects the
appearance of the Healing Grid Illusion. Perception, 41, 243-246.
Marks, L. E. (2004). Cross-modal interaction in speeded classification.
In G. Calvert, C. Spence, & B. Stein (Eds.), The handbook of multi-
sensory processes (pp. 85-105). Cambridge: MIT Press.
Martino, G., & Marks, L. E. (2000). Cross-modal interaction between
vision and touch: The role of synesthetic correspondence. Perception,
29, 745-754.
Miles, L. K., Karpinska, K., Lumsden, J., & Macrae, C. N. (2010). The
meandering mind: Vection and mental time travel. PLoS One, 5,
Palmisano, S., & Chan, A. Y. (2004). Jitter and size effects on vection
are immune to experimental instructions and demands. Perception,
33, 987-1000.
Parise, C. V., & Spence, C. (2012). Audiovisual crossmodal correspon-
dences and sound symbolism: A study using the Implicit Association
Test. Experimental Brain Research, 220, 319-333.
Pilcher, J. J., Nadler, E., & Busch, C. (2002). Effects of hot and cold
temperature exposure on performance: A meta-analytic review. Er-
gonomics, 45, 682-698.
Riecke, B. E., Schulte-Pelkum, J., Caniard, F., & Bülthoff, H. H. (2005).
Influence of auditory cues on the visually-induced self-motion illu-
sion (circular vection) in virtual reality. In M. Slater (Ed.), Proceed-
ings of the 8th International Workshop on Presence 2005 (pp. 49-57).
London: University College London.
Riecke, B. E., Feuereissen, D., & Rieser, J. J. (2008). Auditory self-mo-
tion illusions (“circular vection”) can be facilitated by vibrations and
the potential for actual motion. In S. N. Spence (Ed.), Proceedings of
the 5th Symposium on Applied Perception in Graphics and Visuali-
zation 2008 (pp. 147-54). New York, NY: The Association for Com-
puting Machinery.
Seno, T., Ito, H., & Sunaga, S. (2009). The object and background hy-
pothesis for vection. Vision Research, 49, 2973-2982.
Seno, T., Ito, H., Sunaga, S., & Nakamura, S. (2010). Temporonasal
motion projected on the nasal retina underlies expansion-contraction
asymmetry in vection. Vision Research, 50, 1131-1139.
Seno, T., Ito, H., & Sunaga, S. (2011a). Attentional load inhibits vec-
tion. Attention, Perception, & Psychophysics, 73, 1467-1476.
Seno, T., Ito, H., & Sunaga, S. (2011b). Self-motion perception com-
presses time experienced in return travel. Perception, 40, 497-499.
Seno, T., Ito, H., Sunaga, S., Hasuo, E., Nakajima, Y., & Ogawa, M.
(2011). Consistency between modalities enhances visually induced
self-motion (vection). In G. Suzuki (Ed.), Proceedings of the 12th In-
ternational Multisensory Research Forum. Fukuoka: I-Perception.
Seno, T., Ogawa, M., Ito, H., & Sunaga, S. (2011). Consistent air flow
to the face facilitates vection. Perception, 40, 1237-1240.
Seno, T., Taya, S., Ito, H., & Sunaga, S. (2011). Mental number line in
depth revealed by vection. Perception, 40, 1241-1244.
Seno, T., Yamada, Y., & Ihaya, K. (2011). Narcissistic people cannot
be moved easily by visual stimulation. Perception, 40 , 1390-1392.
Seno, T., & Fukuda, H. (2012). Stimulus meanings alter illusory self-
motion (vection): Experimental examination of the train illusion.
Seeing & Perceiving, 25, 631-645.
Soto-Faraco, S., Spence, C., & Kingstone, A. (2004). Cross-modal dy-
namic capture: Congruency effects in the perception of motion across
sensory modalities. The Journal of Experimental Psychology: Human
Perception and Performance, 30, 330-345.
Williams, L. E., & Bargh, J. A. (2008). Experiencing physical warmth
promotes interpersonal warmth. Science, 322, 606-607.