This work was supported by the European Union Marie Curie Initial Training Network
Marie Curie Actions: FP7-PEOPLE-2013-ITN-604764
Associated URL: http://cognovo.eu
A neuropsycholinguistic chronometric embodied
cognition investigation of the vertical representation of
affect in visuo-spatial target detection
Christopher B. Germann (Ph.D., M.Sc., B.Sc. / Marie Curie Alumnus)
1
Abstract
In two psychophysical experiments we investigated if experimentally induced affective states
(positive vs. negative mood induced by film-clips) bias subsequent performance in a
computerised visual singleton target detection task. Neuropsycholinguistic research indicates
that affective conceptual metaphors consistently associate positive affect with elevated vertical
spatial positions and negative affect with depressed spatial positions. According to the generic
neuronal theory of thought and language, these semantic percept-concept association are
neurocomputationally encoded in Gestalt circuits according to Hebbian principles of long-term
potentiation/depression (LTP/LTD) and spike-timing depended plasticity (STDP), inter alia.
Based on this theoretical background we predicted a priori that positive experimental mood
induction facilitates reaction times (RTs) for target detection in the superior visual field (VF
S
)
relative to the orthogonal inferior visual field (VF
I
). Vice versa, we expected that negative mood
induction facilitates target detection RTs in VF
S
relative to VF
I
(Experiment 1 and 2). In
addition, Experiment 1 explicitly tested the prediction that affective states bias perceptual
judgments on the horizontal axis. Ex hypothesi, we predicted that positive mood induction
facilitates target detection RTs in the right temporal visual field (VF
R
) compared to the
contralateral left field (VF
L
). Per contrast, we predicted the inverse effect for target detection
RTs in VF
L
versus VF
R
. Impetus for the last hypothesis was derived from the
neuropsychological valence model of hemispheric processing of emotion perception which
postulates asymmetric neuronal lateralisation for specific classes of emotions. The data did not
corroborate our directional predictions. All a priori formulated hypotheses were falsified in a
1
The reported experiments were conducted at the Free University of Amsterdam.
Correspondence:
E-Mail: mail@cognovo.christopher-germann.eu
Personal URL: http://cognovo.christopher-germann.eu
Raw data + analysis syntax: http://cognovo.data-sup2.christopher-germann.eu/index.zip
2
quasi-Popperian sense (in support of the urgently needed open-science revolution we provide a
URL to the raw data for the purpose of statistical cross-validation and/or analytical reviews).
Implications of the findings are discussed in the generic theoretical framework of embodied
cognition and we will expound the crucial importance of negative results in a meta-analytic
context (e.g., big-data AI machine learning / publication & confirmation bias). Furthermore,
the de facto irrational epistemological basis of current publishing praxis is criticised from a
universal philosophy of science vantage point. We will close with a critical embodied cognition
discussion of the detrimental effects of constant screen exposure (e.g., ocular fixation on
*smart*phones) on the development of the neuroplastic sensorimotor architecture of children.
Keywords: Embodied cognition, Psychophysics, Affective conceptual metaphors, Spatial
perceptual bias, Dual-process theory, Hemispheric lateralisation, Mood manipulation,
Publication bias, Sensorimotor development.
3
“The words of language, as they are written or spoken, do not seem to play any role in my
mechanism of thought. The psychical entities which seem to serve as elements in thought
are certain signs and more or less clear images which can be “voluntarily” reproduced and
combined. […] The above mentioned elements are, in my case, of visual and some of
muscular type.”
~ Albert Einstein
2
Introduction
The prefatory quotation by Albert Einstein illustrates the involvement of non-linguistic somatic
(physiological) and symbolic/visuo-spatial processes in higher-order abstract thought. The
fundamental question how human beings think about abstract concepts (viz., things they cannot
see, hear, touch, smell, or taste) is deeply rooted in the history of philosophy and hence in the
history of science. The ability to think and communicate about abstract non-tangible domains
such as mathematics, emotions, morality, et cetera, is presumably uniquely human and one of
the hallmarks of human sophistication. Up to date the question how human creatures represent
these abstract meta-physical
3
domains cognitively has not been satisfactorily answered. Earlier
classical cognitive models are based on the dualistic
4
Cartesian assumption of the ontological
disembodiment of mind (or soul,
5
in René Descartes terms; i.e., res extensa versus res
cogitans).
2
As quoted in Hadamard, 1996, The mathematician's mind: The psychology of invention in the mathematical
field. Princeton, NJ: Princeton University Press.
3
Meta-physical is used in the literal sense, i.e., “above physics”. Abstract thoughts do not directly correspond to
physical entities (even though they to use physical experience as a bases, as we shall discuss in greater detail).
4
Baruch de Spinoza formulated a non-dual ontological theory which postulates that mind and body are in
essence two aspect of the same thing (i.c., substance). This is the doctrine of dual-aspect monism which is
compatible with Advaita Vedānta (literally, "not-two", where the prefix “A” is the negation) a philosophical
school of thought which is influential in ancient Indian philosophy (what Aldous Huxley referred to as the
“perennial philosophy”). The negatory approach is based on the principle of “neti neti” transl.: "neither this, nor
that" an epistemological principle which is reminiscent of Popperian falsification, i.e., the hypothetico-
deductive method which is dominant in contemporary science).
5
From a philological vantage point, the word psyche is etymologically derived from the ancient Greek ψυχή
(psukhḗ, “mind, soul, spirit, breath”). According to this definition psychology can thus be regarded as the study
(cf. logos) of the soul. However, today most psychologists are either nescience of this definition or
antagonistically opposed towards it (especially in the contemporary materialistic Zeitgeist, cf. the persistent
influence of Skinnerian behaviourisms and the reductive materialism which dominates much of cognitive
neuroscience). A large proportion of materialistic academic mainstream psychology thereby neglects its
extremely rich intellectual heritage which has deep historical roots in philosophy and which spans many cultures
and epochs (cf. the Vedantic Indian concept “Ātman” (Sanskrit for soul, breath, i.a.) which form the root of the
German word “Atmen” which translates into the English “breath/breathing”)). Psychology is thus the study of
the lifeforce the soul which intimately linked to the breath. It is not a mechanical subject as fields like
computational neuroscience other purely quantitative approaches often suggests even though much of
contemporary terminology is based on the computer metaphor (e.g., the central executive, short-term memory,
dual-process models, etc. pp.) in the same way as much of Freudian theorising was based on thermodynamics.
4
When thought is reintegrated in the bodily matrix (embodied cognition) a very different
perspective on human thinking emerges, namely, that we are not simply inhabitants of our body
we literally utilise it to think through it. The mind-body” problem is a longstanding problem
in the philosophy of mind and the “explanatory gap” is perhaps the most important open
problem contemporary neuroscience. It refers to the following fundamental question: How does
electro-chemical signal transduction in the human brain give rise to thought? In other words,
how do quantitative physical processes (i.e., at the neuronal level or at the subcellular
microtubular level) produce qualitative experiences (e.g., the phenomenological experience of
seeing the colour red). Hitherto, neuroscience and psychology cannot even begin to address this
question in a meaningful way.
6
This quintessential question is closely related to the hard
problem of consciousness.
Hence, the embodied cognition approach has deep philosophical implications. Moreover, it is
not merely of theoretical significance, but it has far-reaching ramifications for computer science
and specifically artificial intelligence and robotics research (e.g., machine learning algorithms
and situated self-learning automata). The broad field of embodied cognitive science is based on
the insight that cognition cannot be understood in vitro. The nucleus of this theoretical
framework is the major premise that cognitive activity is not an isolated phenomenon but that
it is a massively distributed system which spans across the agents physical, social, and cultural
environment. Embodiment and embeddedness are thus crucial factors for contemporary AI
research. It is virtually impossible to formulate a biologically plausible and empirically
adequate theory of intelligent behaviour without reference to (cognitive/neuronal)
representations. A given system is interactively situated in a specific environment which places
characteristic demands on its development (a cybernetic feedback-loop). Organism-
environment interactions are thus essential for the emergence of abstract representations (this
notion is incompatible with the influential Foderianlanguage of thought hypothesis” (LOT)
which stipulates that thought possess a combinatorial syntactical compositional structure
(mentalese) consisting of linguistic tokens, viz., symbol systems which are used for mental
computations, i.e., the computational theory of mind. Somewhat similar to Chomsky’s anti-
Skinnerian universal grammar”, the LOT approach rests on a rationalist model of cognition
6
Perhaps the prima facie assumption that mind and matter are ontological different substances is fallacious.
There are alternative non-dual approaches, but these are quite unpopular in contemporary materialistic science
which assumes that physical matter is primary, and that mind is an emergent property of the former. However,
there is no logical a priori reason for this assumption (it is not supported by empirical evidence) and the inverse
relation is likewise plausible. In fact, cutting-edge evidence from experimental quantum physics points in this
direction (cf. recent refutations of naïve and local realism in the context of violations of Bell-inequalities).
5
and assumes that core aspects of human cognition (specific cognitive modules) are genetically
inherited/innate (i.e., psychological nativism) .
Ergo, the LOT approach stands in sharp contrast with embodied/grounded cognition approaches
(the enduring nature versus nurture debate). From a developmental embodied cognition point
of view, neuronal sensory and motor representations that develop from physical interactions
with the external world (in casu, vertical dimensions) are recycled to assist our thinking about
abstract phenomena. Impetus for this well-developed hypothesis is partially derived from
patterns observed in colloquial language. In order to communicate about abstract concepts,
people regularly utilize metaphors from concrete perceptual domains. As a simple example,
people experiencing {positive} affect are said to be feeling {up} whereas people experiencing
{negative} affect are said to be feeling {down}, viz., affective metaphors are represented in
spatial terms. Cognitive linguists who study cognitive semantics/semiotics (e.g., Gibbs, 1992;
Glucksberg, 2001) have argued that such articulations reveal that people psycholinguistically
conceptualize abstract concepts like affect metaphorically, in terms of physical reality (i.c.,
verticality). It has been argued that without such associative bindings abstract concepts would
lack common ground (e.g., semantic grounding)
7
and it would be difficult (if not impossible)
to convey abstract ideas to other people (Meier & Robinson, 2004). This psycholinguistic
approach enabled interdisciplinary scholars to make important interdisciplinary connections
between embodied experience, neuroanatomy, abstract concepts, and conceptual metaphors.
Embodied Cognition, Conceptual Metaphor Theory, and Hebbian Neuroplasticity
Conceptual Metaphor Theory (Lakoff & Johnson, 1980) defines two basic roles for conceptual
domains posited in conceptual metaphors: 1) the source domain (the conceptual domain from
which metaphorical expressions are drawn) and 2) the target domain (the conceptual domain to
be understood). Conceptual metaphors usually refer to an abstract concept as target and make
use of concrete physical entities as their source. For example, morality is an abstract concept
and when people discuss morality, they recruit metaphors that tap into vertical space (a concrete
physical concept). In colloquial language a person who is moral is often described as ‘‘high
minded’’, whereas an immoral person might be denominated as ‘‘down and dirty’’ (Lakoff &
Johnson, 1999). Following theory the human tendency for categorization is structured by
7
Cf. Ludwig van Wittgenstein’s “language games”, for instance, the “Beetle in a box” Gedankenexperiment
(also known as “the private language argument”; cf. §243 in Wittgenstein’s “Philosophical Investigations”
published in 1953).
6
imagistic, metaphoric, and schematizing abilities that are themselves embedded in the neuro-
biological motor and perceptual infrastructure (Jackson, 1983). Supporters of this view suggest
that cognition, rather than being amodal, is by nature linked to sensation and perception and
consequently inherently cross-modal (e.g., Niedenthal, Barsalou, Winkielman & Krauth-
Gruber, 2005). Furthermore, those researchers argue for the bodily basis of thought and its
continuity beyond the infantile sensorimotor stage (e.g., Seitz, 2000). Cellular Hebbian
mechanism of synaptic strength modulation (long-term potentiation and depression; LTP/LTD)
have been suggested as neuronal-mechanical physiological substrates for conceptual
metaphors. These changes involve processes at the presynaptic level (e.g., direct modulation of
the neurotransmitter release mechanisms and the reuptake machinery) and postsynaptic
processes (modulation of receptor properties, e.g., receptor density and receptor morphology),
inter alia.
8
Developmental neuro/synapto-plasticity thus likely plays a crucial role in this
context (i.e., theories of synaptic sprouting & pruning). Indeed, a large corpus of research
suggest that the neurological processes that make abstract thought possible are intimately
connected with the neurological processes that are responsible for representing perceptual
experiences (we call this the “neuronal perception-cognition continuum”; in other words, a
neuroanatomical coupling of perceptual and higher cognitive functions). In line with this
argument, it appears that the motor system (i.e., hierarchically organised topographic maps in
the parietal cortex which represent the sensory organs) plays a central neuroanatomical rôle in
higher-order abstract metaphorical thinking (the mirror-neuron system and the mental
simulation of behavioural actions are crucial components in this context). This correlation is
assumed to be unidirectional. Specifically, it is argued that conceptual thought is based on
sensory experience, but sensory experience is not based on conceptual thought (e.g., love is a
rose, but a rose is a rose; Meier & Robinson, 2005).
9
Interestingly, a similar explanatory framework forms the basis contemporary dual-process
theories of cognition. For instance, Kahneman (2003) argues that higher-order cognitive
8
A rapidly growing body of research indicates that non-synaptic cellular plasticity mechanisms accompany
synaptic changes (presumably concomitantly), for instance, modulations of neuronal excitability (viz., alterations
of voltage-dependent membrane conductance via voltage-gated channels). The functional significance of these
processes remains largely elusive, but they might play an important role in Hebbian learning. Further, it is
important to emphasise that science has no theory how mental contest (ideas) “emerge” from physical properties
(neurons) this is the perennial “explanatory gapin neuroscience and philosophy of mind.
9
According to theory, this unidirectional correlation I based on time-dependent plasticity at the level of neuronal
pairing (first the perceptual source domain is activated and then the abstract target domain). That is, first neuron
A which is responsible for perception is activated which is then followed by neuronal firing in neuron B
(spreading activation is unidirectional; viz., paired stimulation for spike-timing-dependent plasticity always
follows the direction A => B).
7
processes are conditioned on the principles of lower-level perceptual processes. He advocates
an evolutionary perspective on abstract reasoning and his reflections are based on the
assumption that there is a kind of quasi biogenetic progression in the evolution of cognitive
processes, starting from automatic perceptual processes which form the fundamental basis for
the evolution of more deliberate and abstract modes of information processing. The postulated
phylogenetic history of cognitive processes can be adumbrated as follows:
PERCEPTION => INTUITION => REASONING
According to this sequential view on the Darwinian evolution of cognitive systems, perception
appears early on the timeline of history (and ergo present in many species), whereas abstract
higher-order reasoning evolved relatively recently (a hallmark of human cognition). Intuition
is intermediate between the automatic (System 1) processes of perception and the deliberate,
higher-order reasoning (System 2) processes that are the hallmark of human intelligence
(Kahneman, 2003).
This line of reasoning connects neatly with the embodied cognition framework expounded
above. The impetus behind the present research is thus essentially based on the following
question: Why is an abstract concept like affect so frequently linked to concrete perceptual
qualities like vertical position? And what are the specific neuro-cognitive mechanisms which
undergird this association?
One possible explanation for this perceptual-conceptual” connection comes from classical
developmental research. Early theorists of sensorimotor learning and development emphasized
the importance of movement in cognitive development (e.g., Piaget, 1952). According to this
perspective, human cognition develops through sensorimotor experiences. Young children in
the Piagetian sensorimotor stage (from birth to about age two) think and reason about things
that they can see, hear, touch, smell or taste. Motor skills emerge and the infant cultivates the
coordination of tactile and visual information. Later developmental researchers postulated that
thinking is an extended form of these more fundamental behaviours and that it is based on these
earlier modes of adaptation to the physical environment (Bartlett, 1958). For example, it has
been suggested that gesture and speech form parallel systems (McNeill, 1992) and that the body
is central to mathematical comprehension (Lakoff & Nunez, 1997).
Thus, when children grow older, they develop the skills to think in abstract terms and these
higher skills are built upon more primitive earlier sensorimotor representations. For example, a
8
warm bath leads to a pleasant sensory experience and positive affect (thalamic/limbic
processes). In adulthood, this pairing of sensory and abstract representations may give rise to a
physical metaphor (e.g., a warm person is a pleasant person) that continues to exert effects on
representation and evaluation (Meier & Robinson, 2004).
10
Transferred to the vertical
representation of affect one can currently only speculate. Tolaas (1991) proposes that infants
spend much of their time lying on their back. Rewarding sensory stimuli like food and affection
arrive from a high vertical position. The caregiver frequently appears in the infant’s upper
visual-spatial environment (Meier, Sellbom & Wygant, 2007). Given that neuroplasticity is still
very high in the developing brain, specific associative neuronal pathways are formed by
repetition (via LTP/LTD/STDP). As children age, they use (recycle) this sensorimotor
foundation to develop abstract thought, as recognized by developmental psychologists (e.g.,
Piaget & Inhelder, 1969). This early conditioning causes adults to use the vertical dimension
when expressing and representing affect. These considerations suggest that the link between
affect and vertical position may develop early in the sensorimotor stage (see Gibbs, 2006; for
sophisticated considerations) due to specific neuroplastic adaptations in brains circuitry
11
which
are primarily based on experience (e.g., sensory input). What follows are some paradigmatic
exemplars of primary metaphors in the spatial domain (orientational metaphors) which couple
the prelinguistic perceptual source domain with the abstract conceptual target domain:
Up => More
Down => Less
Up => Happy
Down => Sad
Up =>Moral
Down => Immoral
N.B. These association can vary across cultures (i.e., language-specific conceptualisations are
at variance in different linguistic milieus).
10
Another example would be the involvement of the insular cortex in disgust be it “olfactory disgust”,
“gustatory disgust” or “moral disgust”.
11
In addition to neuroplastic changes, synaptic pruning plays a formative role in the formation of experience-
specific neurocircuitry. Especially during approximately the first five years of development structural and
functional brain networks “crystallise”. This system-level rewiring constantly defines the connectivity of the
developing brain in an experience-dependent manner and it is assumed that a large amount of neurons is pruned
during these developmental stages (but see Ablitz, et al., 2007).
9
The classical “Aristotelian Peripatetic Axiom” is of pertinence in this context. This
epistemological argument highlights the importance of sensory input in the context of reasoning
and knowledge:
Nihil est in intellectu quod non sit prius in sensu.
Translation from the Latin: Nothing is in the intellect that was not first in the senses.
According to this empirical stance, the human mind is furnished by sensory impressions which
in turn form the basis of ideas. In other words, the data which impinges on our sensorium
“determines” our thoughts (and, in sensu lato, our Weltanschauung/worldview).
John Locke used a similar line of reasoning in his classic text “An Essay Concerning Human
Understanding” (1690), specifically his chapter “On the Association of Ideas”.
12
According to Locke, empirical (sensory) data which is accumulated diachronically (during the
course of a lifetime) forms the basis of thought. Locke argues that the most elementary act is
the sensory act and the most elementary contents of the mind are sensations. He remarks: For
to imprint anything on the mind without the mind's perceiving it, seems to me hardly
intelligible” (op. cit., Chapter II - On Innate Ideas). In other words, what enters the mind comes
through the sensorium and these elementary sensations are then connected. According to
Newton, the corpuscular components of reality are held together by gravitational forces, i.e.,
per analogiam with Newton's law of universal gravitation which follows the inverse-square
law.
13
Locke ingeniously transferred the Newtonian idea of physical gravitation to elementary
sensations and proposes the principle of mental association as the psychological counterpart.
14
Events which are experienced together frequently are connected by associative processes.
15
12
The full text of this timeless classic is available under the following URL:
https://archive.org/details/essay_concerning_human_understanding2_1904_librivox/humanunderstanding
13
The inverse-square law can be mathematically notated as follows: gravitational intensity
1
distance
2
In the context of Locke’s psychological theory, the term “gravitational intensity” can be replaced with
“associational intensity”. While gravitation is the attraction of two physical objects, association describes the
attraction between mental concepts (i.e., ideas). For instance, the “distance” between various concepts can be
indirectly quantified by variations in reaction-times in a semantic priming paradigm, for instance, the implicit-
association test (IAT) (Greenwald & Farnham, 2000; Sriram & Greenwald, 2009). The concepts “tree” and “roots”
are closer associated (i.e., the “associational intensity” is stronger) than the concepts “university” and “beer”
(perhaps this is an unfortunate example, but it illustrates the general point). Locke thus combined physics with
psychology and he could thus perhaps be regard as one of the first psychophysicist (long before Fechner).
14
Interestingly, it has been noted by historians of philosophy and science that “Locke's attitude towards the
nature of ideas in the Essay is reminiscent of Boyle's diffident attitude towards the nature of matter” (Allen,
2010, p. 236).
15
This Lockean idea can be regarded as the predecessor of Hebbian engrams and assembly theory “cells that
fire together wire together” (Hebb, 1949).
10
These then recombine to form simple ideas. Out of simple ideas, increasingly complex ideas
are hierarchically assembled by the binding force of association this the Lockean associative
“logic of ideas”. The Lockean associationist memetic
16
account is de facto of great pertinence
today (e.g., machine learning algorithms, associative (Bayesian) neural networks in artificial
intelligence, deepmulti-layered convolutional networks, etc.). Locke was clearly far ahead
of his time.
In the context at hand, Lockean associationism can be regarded as the predecessor of Hebbian
engrams and cell assembly theory which function according to the Hebbian mantra “cells
that fire together wire together” (Hebb, 1949). There is general consensus among embodied
cognition researchers that Hebbian principles form the neuronal basis of conceptual metaphors
(but see Lakoff, 2014). The ancient empiricist axiom of how the mind is furnished through the
sensorium has thus been corroborated by fundamental research in 20
th
century neuroscience
and cybernetics. According to Hebbian theory connected cells combine into engrams. Multiple
cells, in turn, become cell assemblies (cf. Locke’s description above). Hebb stated the following
to describe his approach:
“The general idea is an old one, that any two cells or systems of cells that are repeatedly active
at the same time will tend to become ‘associated’ so that activity in one facilitates activity in
the other.” (Hebb, 1949, p.70)
Hebb wrote further:
“When one cell repeatedly assists in firing another, the axon of the first cell develops synaptic
knobs (or enlarges them if they already exist) in contact with the soma of the second cell.” (op.
cit. p.63)
A conscience mathematical formulaic description of Hebb's postulate is provided by the
following formula:

=
1
=1
,
where w
ij
is the weight of the connection from neuron j to neuron i, p is the number of training
patterns, and x
i
k
the k
th
input for neuron i.
16
The science of memetics tries to (mathematically) understand the evolution of memes, analogous to the way
genetics aims to understand the evolution of genes (Kendal & Laland, 2000). Locke’s early contributions are
pivotal for the development of this discipline which is embedded in the general framework of complex systems
theory (Heylighen & Chielens, 2008). Memetics is of great importance for our understanding of creativity and
the longitudinal evolution of ideas in general. Memes reproduce, recombine, mutate, compete and only the best
adapted survive in a given fitness landscape. Similar to genotypes, the degree of similarity/diversity between
memes (and their associated fitness values) determines the topology of the fitness landscape.
11
Thus, repetition of specific information patterns consolidates specific information processing
pathways in the brain (per analogiam to a path in a forest which becomes more defined and
more widely used the more people walk people walk on it, similarly, the ocean washes specific
grooves into cliffs over elongated periods of time). Accordingly, the Weltanschauung
(worldview) of human beings is shaped by the information we are exposed to. The crux of this
line of reasoning is that the brain learns by repetition and that sensory information shapes our
brain and hence our mental models of reality. It can thus be agued that conceptual metaphors
are based on the same Aristotelian/Lockeian/Hebbian principle of association between
sensation and thought. It is the pairing of sensory inputs which creates the associations (i.e.,
neuro-computational coupling / quasi “gravitational” binding) between the source domain and
the target domain of conceptual metaphors. These circuits are not restricted to specific
modalities (amodal circuitry) and based on pathways that cascade across sensory maps to create
neuronal bindings (Gestalts circuits) which form the basis of conceptual metaphors which
structure language and thought.
From theory to praxis: Experimental applications of conceptual metaphor theory
Affective metaphor theory has recently gained a lot of attention in the domain of political
decision-making (Hammack & Pilecki, 2012; Lin & Chiang, 2015; Miller, 2006; Musolff,
2004). Such unconscious association can for instance be utilised for the purpose of impression-
management as “persuasive devices” (Mio, 1997), particularly in the context of public relations
à la Bernays (cf. Schölzel, 2014). Affective metaphors and related metaphoric associations
apply to a multitude of perceptual dimensions such as, for example, spatial location, brightness
and tone pitch. A plethora of studies investigated the link between abstract concepts (i.c., affect)
and physical representation (i.c., verticality). For example, in a study conducted by Meier and
Robinson (2004) participants had to evaluate positive and negative words either above or below
a central cue. Evaluations of negative words were faster when words were in the down rather
than the up position, whereas evaluations of positive words were faster when words were in the
up rather than the down position. In a second study, using a sequential priming paradigm, they
showed that evaluations activate spatial attention. Positive word evaluations reduced reaction
times for stimuli presented in higher areas of visual space, whereas negative word evaluations
reduced reaction times for stimuli presented in lower areas of visual space. A third study
12
revealed that spatial positions do not activate evaluations (e.g., “down” does not activate
‘‘bad’’). Their studies give credit to the assumption that affect has a physical basis.
Moreover, a seminal study by Wapner, Werner, and Krus (1957) examined the effects of
success and failure on verticality related judgements. They found that positive mood states,
compared to negative mood states, were associated with line bisections that were higher within
vertical space.
In a more recent study Meier et al. (2007) reported that people have implicit associations
between God-Devil and up-down. Their experiments showed that people encode God-related
concepts faster if presented in a high (vs. low) vertical position. Moreover, they found that
people estimated strangers as more likely to believe in God when their images appeared in a
high versus low vertical position.
Another study by Meier and Robinson (2006) correlated individual differences in emotional
experience (neuroticism and depression) with reaction times with regard to high (vs. low)
spatial probes. The higher the neuroticism or depression of participants, the faster they
responded to lower (in contrast to higher) spatial probes. Their results indicate that negative
affect influences covert attention in a direction that favours lower regions of visual space. In
second experiment the researchers differentiated between neuroticism and depression. They
argued that neuroticism is more trait-like in nature than depression (which is more state-like).
The researchers concluded from their analysis that depressive symptoms were a stronger
predictor of metaphor consistent vertical selective attention than neuroticism.
Similar results emerged when dominance-submission was assessed as an individual difference
variable and a covert spatial attention tasks was used to assess biases in vertical selective
attention (Robinson, Zabelina, Ode & Moeller, in press). Linking higher levels of dominance
to higher levels of perceptual verticality they found that dominant individuals were faster to
respond to higher spatial stimuli, whereas submissive individuals were faster to respond to
lower spatial stimuli.
Further support for the Conceptual Metaphor Theory comes from a study investing the extent
to which verticality is used when encoding moral concepts (Meier, Sellbom & Wygant, 2007).
Using a modified IAT
1
the researchers showed that people use vertical dimensions when
processing moral-related concepts and that psychopathy moderates this effect.
Inspired by the observation that people often use metaphors that make use of vertical positions
when they communicate concepts like control and power (e.g. top manager vs. subordinate),
some researchers investigated social structure from a social embodiment perspective. For
13
example, Giessner and Schubert (2007) argued that thinking about power involves mental
simulation of vertical location. The researchers reported that the description of a powerful
leader led participants to place the picture of the leader in an organization chart significantly
higher as compared to the description of a non-powerful leader.
As mentioned above, affective metaphors and related associations apply multitudinous
perceptual dimensions. Recent research examined the association between stimulus brightness
and affect (Meier, Robinson & Clore, 2004). The investigators hypothecated that people
automatically infer that bright things are good, whereas dark things are bad (e.g., light of my
life, dark times). The researchers found that categorization was inhibited when there was a
mismatch between stimulus brightness (white vs. black font) and word valence (positive vs.
negative). Negative words were evaluated faster and more accurately when presented in a black
font, whereas positive words were evaluated faster and more accurately when presented in a
white font. In addition, their research revealed the obligatory nature of this connection.
Furthermore, a series of studies showed that positive word evaluations biased subsequent tone
judgment in the direction of high-pitch tones, whereas participants evaluated the same tone as
lower in pitch when they evaluated negative words before (Weger, Meier, Robinson & Inhoff,
2007).
In addition, recent experimental work supports the notion that experiences in a concrete domain
influence thought about time (an abstract concept). Researchers assume that, in the English
language, two prevailing spatial metaphors are used to sequence events in time (e.g., Lakoff &
Johnson, 1980). The first is the ego-moving metaphor, in which the observer progresses along
a timeline toward the future. The second is the time-moving metaphor, in which “a time-line is
conceived as a river or a conveyor belt on which events are moving from the future to the past”
(Boroditsky, 2000, p. 5). In an experimental study by Boroditsky and Ramscar (2002),
participants had to answer the plurivalent question: “Next Wednesdays meeting has been moved
forward two days. What day is the meeting now that it has been rescheduled?” Before asking
this ambiguous question participants were led to think about themselves or another object
moving through space. If participants were led to think about themselves as moving forward
(ego-moving perspective), then participants answered more often “Friday”. On the other hand,
if had thought of an object as moving toward themselves (time-moving perspective), then they
more often answered “Monday”. The researchers showed that those effects do not depend on
linguistic priming, per se. They asked the same ambivalent question to people in airports.
14
People who had just left their plane responded more often with Friday than people who were
waiting for someone.
Moreover, cognitive psychologists have shown that people employ association between
numbers and space. For example, a by study Dehaene, Dupoux and Mehler (1990) showed that
probe numbers smaller than a given reference number were responded to faster with the left
hand than with the right hand and vice versa. These results indicated spatial coding of numbers
on mental digit line. Dehaene, Bossini and Giraux (1993) termed the mentioned association of
numbers with spatial left-right response coordinates the SNARC-effect (Spatial-Numerical
Association of Response Codes). Another SNARC-effect related issue is that empirical data
indicates that associations between negative numbers with left space exist. For example, in a
study by Fischer, Warlop, Hill and Fias (2004) participants had to select the larger number
compared to a variable reference number of a pair of numbers ranging from 9 to 9. The results
showed that negative numbers were associated with left responses and positive numbers with
right responses. The mentioned results support the idea that spatial association give access to
the abstract representation of numbers. As mentioned above, master mathematicians like
Einstein explicitly accentuate the role of the concrete spatial representation of numbers for the
development of their mathematical ideas. Today there are a few savants which can do
calculation up to 100 decimal places. They also emphasize visuo-spatial imagery as in the case
of Daniel Tammet
2
who has an extraordinary form of synæsthesia which enables him to
visualize numbers in a landscape and to solve huge calculations in the head. Moreover, about
15% of ordinary adults report some form of visuo-spatial representation of numbers (Seron,
Pesenti, Noel, Deloche & Cornet, 1992). This implies that the integration of numbers into visuo-
spatial coordinates is not a rare phenomenon.
The mentioned studies show that abstract concepts (invisible and intangible) have an
astonishing physical basis and that many dimensions of the physical world help us to represent
those abstract domains. In addition to the hypothecated vertical representation of affect we
wanted to explore if positive mood facilitates target detection in the right visual field compared
to target detection in the left visual field and vice versa. Perhaps it is possible that there is a
connection between the SNARC effect and the horizontal perceptual bias we hypothecate. It
might be speculated that there is a connection between negative and small numbers which are
apparently associated with left responses and negative affective states (which could facilitate
target perception in the left visual field). In both cases the right (contra lateral) hemisphere is
engaged. For example, the valence model of hemispheric processing of emotion perception
15
assigns specific functions to each hemisphere. The following paragraph tries to shed some light
on the question how people perceive and process emotional stimuli from a neuropsychological
point of view.
The neuropsychological valence model of hemispheric processing of emotion
perception
The widely debated neuropsychological valence hypothesis assumes that the right hemisphere
is specialized for negative emotion and that the left hemisphere is specialized for positive
emotion (Ehrlichman, 1987; Silberman & Weingartner, 1986). Many studies and in particularly
those concerned with the perception of emotion report hemispheric differences as a function of
positive vs. negative emotions (e.g., Sato, Kochiyama, Yoshikawa, Naito & Matsumura, 2004;
Van Strien & Luipen, 1999; Heller & Nitschke, 1997; Adolphs, Jansari & Tranel, 2001).
Patients who had an injury in the right hemisphere are more likely to have trouble perceiving
negative emotions in affective faces but the perception of posititive emotions is typically
preserved (Adolphs, Damasio, Tranel & Damasio, 1996). Further support for the relative
trueness of the valence hypothesis comes from studies recording event-related potentials
(ERPs) reporting that emotional responses to aversive stimuli facilitated right hemisphere
processing during higher cognitive tasks (Simon-thomas, Role & Knight, 2005).
Countless amytal injection studies found that injection at the left carotid artery induced extreme
reactions, which were characterized by guilt, crying, pessimistic statements and worries about
the future (Rossi & Rosadini, 1967; Silberman & Weingartner, 1986). On the other hand,
injection at the right carotid artery caused an euphoric reaction that consisted of lack of
apprehension, laughing, smiling and a sense of well-being. The mood changes associated with
the injection of one hemisphere are thought to represent the release of one hemisphere from
contra lateral inhibitory influences (Silberman & Weingartner, 1986). These findings have been
supported by observations of patients suffering from brain damage in which individuals with
left-hemisphere lesions were more likely to show “catastrophic” behaviour, whereas patients
with right hemisphere lesions were more likely to show “indifferent” behaviour (Heilman et al.,
1975).
Early tachistoscopic studies provided supporting results for the valence hypothesis. For
example, Reuter-Lorenz, Givis and Moscovitch (1983) presented happy or sad facial
expressions in one visual field while simultaneously presenting a neutral expression in the other
16
visual field. Participants were instructed to identify the side that contained the emotional face.
Reaction times were shorter for happy faces shown within the right visual field relative to happy
faces shown within the left visual field. Moreover, sad faces shown within the left visual field
where processed faster compared to sad faces shown within the right visual field. Subsequent
studies also found that in some cases, negative affective faces are identified more rapidly or
more accurately when presented within the left visual field (e.g., Everhart & Harrison, 2000).
Perhaps the most persuasive results in support of the valence hypothesis were derived from a
vast number of EEG studies that have linked increased left hemisphere activity with positive
affect and increased right-hemisphere activity with negative affect. For example, Davidson et
al. (1979) instructed participants to indicate their emotional responses while watching a
television program with varying emotional content. EEG measurements displayed relative left-
hemisphere activity during a positive emotional response, whereas the opposite pattern was
observed during a negative emotional response. However, it should be noted that other EEG
researchers testing the valence hypothesis have sometimes reported the opposite pattern of
activation (e.g., Schellberg, Besthern, Pfleger & Gasser, 1993).
A more differentiated version of the valence hypothesis (Davidson, 1984) suggests that the
frontal regions of both hemispheres are involved in the experience and expression of emotions
whereas the posterior regions of the right hemisphere are involved in the perception of emotion.
Moreover, frontal regions of the left hemisphere are specialised for positive emotions and
frontal regions of the right hemisphere are specialised for negative emotions.
A priori predictions
In accord with the mentioned theories (conceptual metaphor theory, valence model of
hemispheric processing of emotion perception) we predicted that positive and negative mood
states cause systematic differences in visuo-spatial information processing. Functionalistic
perspectives on affect act on the assumption that affect has an informative value and that cues
from positive and negative emotions/mood may be experienced as information that promotes
attention to global and local information, respectively (Gasper & Clore, 2002). Our primary
hypothesis predicted that being in a positive or negative mood influences information
processing in a metaphorically consistent directional manner. We expected that participants in
a positive mood respond faster to targets presented to the superior visual field (relative to targets
presented to the inferior visual field) and that negative mood facilitates target detection in the
17
inferior visual field (relative to targets presented to the superior visual field). We investigated
this hypothesis in Experiment 1 and 2. Furthermore, our secondary hypothesis forecasted that
participants in a positive mood will show shorter reaction times with regard to targets presented
to the right visual field compared to targets presented to the left visual field, whereas
participants in a negative mood should show the opposite pattern. The last hypothesis is in line
with the valence model of hemispheric processing of emotion.
Experiment 1
We conducted Experiment 1 to determine if positive mood facilitates target detection in the
superior visual field relative to target detection in the inferior visual field. Consequently, we
were interested whether negative mood facilitates target detection in the inferior visual field
relative to target detection in the superior visual field. In addition, we aimed to examine if
positive mood facilitates target detection in the right visual field compared to target detection
in the left visual field, whereas negative mood facilitates target detection in the left visual field
compared to target detection in the right visual field.
Method
Participants and Design
Participants were 48 students at the Vrije Universiteit Amsterdam which participated in this
study on a voluntary basis and were naïve to the purpose of the study. They either received
course credit or participated on a paid basis. The study employed a mixed design. The between-
participants independent variable was the mood induction procedure with two levels (positive
vs. negative mood induction) and the within-participants dependent variable was the reaction
time on four different target locations (up, down, left, right) in a visual search task.
Materials and procedure
Each participant was seated in a sound-attenuated, 2x2x2m cubicle (amplisilent). A high-
resolution color computer monitor (800x600 pixel screen-resolution, 24-bit truecolor color
depth, 59 Hertz refresh rate) was located at eye level. They were informed that all instructions
and tasks would be presented on screen.
18
Mood induction procedure
Participants were randomly assigned to one of two different mood induction conditions
(positive vs. negative mood). They viewed a film clip, which lasted approximately 7min in both
conditions. The positive film clip was a scene from “Jungle Book,” and the negative clip was a
scene from the film “Sophie’s choice” (see Appendix for details). Meta-analytic research
indicates that film clips are an effective and reliable method to induce both positive and negative
mood (for a systematic comparison of various mood-induction procedures see Westermann,
Spies & Hesse, 1996).
Mood manipulation check
Directly after the mood induction participants were asked to indicate how they ‘feel at this
moment’ which they did on a modified four-item Visual Analog Scale (Aitken, 1969), ranging
from 0=not at all to 100=very much. The four items participants responded to were: I feel
positive; I feel negative; I feel happy; I feel sad. Participants were instructed to indicate their
mood by mouse clicking on a position of the Visual Analog Mood Scales (VAMS; consisting
of a single, horizontal white line on a black background) that best represented their affective
state. Furthermore, participants responded to a control item (I feel fraught) to ensure that they
understood the task. The four Visual Analog Mood Scales were presented in succession and in
a counterbalanced order. Prior research indicated that the VAMS is a brief, easily administered,
and valid measures of mood states (Stern, 1997). For example, the VAMS correlated
significantly with other, more extensive and verbally demanding mood measures like the POMS
Depression/Dejection scale and the Beck Depression Inventory.
19
Visual search task.
Subsequent to the mood manipulation check, participants completed the visual search task
which was presented on the same computer-monitor. Stimuli were presented using E-Prime
software (Psychology Software Tools, Inc.; Pittsburgh, PA). A white fixation dot (0.15° of
diameter at an observation distance of ≈75cm) was displayed on a black background in the
centre of the computer screen throughout the experiment. Participants were instructed to focus
on the fixation dot and not to move the eyes during the course of any trial (steady ocular
fixation). It was emphasized that steady fixation would facilitate fast and accurate reactions
(i.e., speed-accuracy trade-offs). The search display consisted of four grey outline circles
(approximately 1.5° of diameter) on a black background. The midpoints of the outline circles
superposed the vertexes of an imaginary square diamond shape (6.84° x 6.84°) and the fixation
dot constituted the centre of this figure. When a target was present all the outline circles were
grey except for one, which constituted the target and was either red or green. Half of the trials
the target was present, whereas in the other half the target was absent. In target-present trials
the target position was randomized among the four possible element positions. Moreover, half
of the target-present trials consisted of a green target, whereas the other half consisted of a red
target. Participants were instructed to indicate as fast and accurate as possible whether the target
was present or absent and to press the appropriate response key with one of her/his index fingers
resting on the response keys. The response panel consisted of a standard ASCII keyboard and
the response keys were the “m” and “z” keys. Furthermore, the order of these keys was
counterbalanced so that half of the participants indicated “yes” with their right hand and the
other half indicated “yes” with their left hand. Each participant performed a practice block
consisting of 32 trials and after that a total of 5 experimental blocks, each with 32 trials (a total
of 160 experimental trials). Participants received feedback regarding their reaction time (in
milliseconds) and accuracy (in percentage) after each block and were asked to write down this
feedback on a spreadsheet which was laid out in front of them. If the mean accuracy of the
practice block was lower than 60% participants had to repeat the 32 practice trials.
The sequence of events was as follows: Initially, the white fixation dot was presented for
750ms, followed by the search display. The search display remained present for a maximum of
2sec until a response was emitted. In case no response was made after 2sec, the participants
were requested to “React faster!” (a salient message in red font appeared). In case of a false
response the participants were informed by the notification “Error!” (a message in red font).
20
Figure 1 shows the various display configurations in a trial sequence. Upon completion of the
visual search task, the participants were debriefed about the objective of the study. Finally,
participants were thanked for their time, remunerated and released.
Figure 1. An example of a trial sequence in the search task. The fixation dot was presented for
750ms, followed by the stimulus display. Percipients had two seconds to respond.
Results of Experiment 1
Validation of experimental mood manipulations
To examine the effectiveness of the mood induction procedure an independent-samples t-test
was conducted to compare self-reported mood between mood induction conditions. Participants
indicated their mood on a four item Visual Analog Mood Scale which was scored from 0 to
100, based on the distance in millimetres from the left pole. Two VAMS items which measured
negative affect (I feel negative; I feel sad) were reverse coded so that high ratings indicated
positive mood. The four VAMS items formed an internally consistent scale (Cronbach’s α =
21
.98), therefore we analysed the mean composite score of the four items (George & Mallery,
2003). The t-test was statistically significant, t(46) = 14.83, p < .05, η2 = .83, indicating that
participants in the negative mood induction condition (M = 25.42, SD = 15.54) on the average
reported less positive feelings than those in the positive mood-induction condition (M = 84.65,
SD = 11.88). The eta square index indicated that 83% of the variance of the dependent variable
(mean composite mood score) was accounted for by whether a participant was assigned to the
positive mood induction condition or the negative mood induction condition.
Statistical reaction time analysis
Only trials in which the target was present were included in the statistical analysis. Furthermore,
inaccurate trials were dropped from the analysis. The total percentage of errors in response to
the target was 5.57%. Reaction times were then log-transformed to reduce the skewness of the
distribution. Log transformation tends to normalize distributions and reduces the impact of long
response times in the tails of the distributions, therefore reducing the impact of long outliers,
leading to higher power for ANOVA (Ratcliff, 1993). Trials that were 2.5 standard deviations
below or above the grand mean were excluded from the analysis, as these likely reflect
erroneous responses or lapses of attention.
A 2 (mood induction condition: positive vs. negative) x 2 (target position: top vs. bottom)
repeated measures ANOVA was performed on the log-transformed reaction times.
17
The
between subjects’ factor was the mood induction condition and the within subject factor was
the target position. Neither of the main effects was significant: F(1, 46) = 2.94, p > .05, η2 =
.06, for the mood induction condition and F < 1, for the target position. Contrary to our a priori
predictions, the mood induction x target position interaction was also nonsignificant, F < 1. The
results are visualised in Figure 2.
Furthermore, a 2 (mood induction condition: positive vs. negative) x 2 (target position: left vs.
right) repeated measures ANOVA was conducted. The main effects were nonsignificant: F(1,
46) = 3.58, p > .05, η2 = .06, for the mood induction condition and F < 1, for the target position.
Contrary to our prediction, the mood induction x target position interaction was also
nonsignificant, F < 1.
17
N.B.: We used “conventional” Type III sum of squares (SS) in all reported analyses. It should be noted that it
is strange that for regression (forward, backward, stepwise methods), the type of SS is always reported while
thuis information is usually neglected for ANOVA, despite the fact that it is based on the same linear model (but
see Howell, 2002).
22
Statistical error rate analysis
The total percentage of errors in response to the target was 5.57%. In addition, errors rates were
calculated for each condition for each participant. A repeated-measures ANOVA was
performed on these totals with target location (top vs. bottom) as the within subject factor and
mood induction condition as the between subject factor. The factor target location was not
significant, F < 1. There was no significant effect of mood induction condition, F(1, 46) = 2.50,
p > .05, and the interaction between target location and mood induction condition was also
nonsignificant, F(1, 46) = 3.78, p > .05. The results mimicked those found in the reaction time
data. Furthermore, a repeated-measures ANOVA was performed on the target miss rates with
target location (left vs. right) as the within subject factor and mood induction condition as the
between subject factor. The factor target location was not significant, F < 1. There was no
significant effect of mood induction condition, F < 1, and the interaction between target location
and mood induction condition was also nonsignificant, F < 1. This pattern was also present in
the reaction time data.
Figure 2. Mean reaction time for target present trials as a function of mood induction condition.
23
Discussion of Experiment 1
The results of our analysis confirmed that the mood induction procedure was effective.
Participants in the positive mood induction condition on the average reported more positive
feelings than those in the negative mood induction condition. However, the ANOVA results of
the reaction time analysis did not support our hypotheses. Mood did not facilitate visual
attention in any direction (neither vertical nor horizontal). The error rate analysis was congruent
with the RT analysis.
Experiment 2
Experiment 2 was basically a reductive variation of Experiment 1. Given the nonsignificant
results of Experiment 1 we tried to eliminate potentially confounding factors and focused our
interest exclusively on the vertical representation of affect. We chose to conduct a visual
discrimination task to get a clearer picture of selective visual attention as a function of affect
(i.e., positive vs. negative mood). It is thinkable that the stimuli in Experiment 1 activated the
whole visual field because targets were presented to the upper, lower, left and right visual field.
It is plausible that this activation masked the hypothesised perceptual bias on the vertical
dimension. To address this potential confound methodologically we conducted a second
experiment where the target was located either at the top or the bottom of the stimulus display.
Method
Participants and design
Participants were 20 students at the Vrije Universiteit Amsterdam and participated in this study
on a voluntary basis. They either received course credit or took part on a paid basis. The study
used a mixed design. As in Experiment 1, the between-participants independent variable was
the mood induction procedure with two levels (positive vs. negative mood induction) and the
within-participants dependent variable was the reaction time on two different target locations
(top vs. bottom) in a visual discrimination task.
24
Materials and procedure
In order to facilitate procedural commensurability, the stimulus field and the equipment were
largely identical to Experiment 1. Per contrast to Experiment 1 the target position was
randomized among two (instead of four) possible element positions from trial to trial (the target
appeared either at the top or the bottom of the computer screen). Furthermore, the target was
always present. In order to maximise the interspace between the target stimuli we reduced the
diameter of the outline circles to approximately 0.75° of diameter. The instructions were the
same as in Experiment 1.
The mood induction procedure and the subsequent validation of its effectiveness were identical
to Experiment 1.
Visual search task.
Directly after the mood manipulation check participants completed the visual discrimination
task. The target was presented either above or below the central fixation dot (in a randomized
fashion). The task of the participants was to discriminate whether the target appeared above or
below the central fixation dot. As in Experiment 1, the response keys were the “m” and “z”
keys. The order of these keys was counterbalanced so that half of the participants indicated
above” with their right hand and the other half indicated “above” with their left hand.
Consequently, half of the participants indicated “below” with their right hand and the other half
indicated “below” with their left hand. The instructions, the arrangement of trial-blocks and the
feedback were similar to Experiment 1. Figure 3 illustrates the various display configurations
in a trial sequence.
25
Figure 3. An example of a trial sequence in the discrimination task. The fixation dot was
presented for 750ms, followed by the stimulus display. Percipients had to respond within 2
seconds.
Results of Experiment 2
Validation of experimental mood manipulations
As in Study 1, an independent-samples t-test was carried out to compare self-reported mood
between mood induction conditions. The four VAMS items formed an internally consistent
scale (Cronbach’s α = .92), therefore we analysed the mean composite score of the four items.
The t-test was significant, t(18) = 4.65, p < .05, η2 = .55. Participants in the negative mood
induction condition (M = 29.85, SD = 18.56) on the average reported fewer positive feelings
than those in the positive mood-induction condition (M = 71.70, SD = 21.60). The eta square
index indicated that 55% of the variance of the dependent variable was accounted for by
26
whether a participant was assigned to the positive mood induction condition or the negative
mood induction condition.
Statistical reaction time analysis
As in Study 1, inaccurate trials were dropped from the analysis. The total percentage of errors
in response to the target was 7.31%. Reaction times were then log-transformed. Trials that were
2.5 standard deviations below or above the grand mean were excluded from the analysis. A 2
(mood induction condition: positive vs. negative) x 2 (target position: top vs. bottom) repeated
measures ANOVA was performed on the transformed reaction times. The between subjects’
factor was the mood induction condition and the within subject factor was the target position.
There were no significant main effects: F(1,18) < 1, for the mood induction condition and F(1,
18) = 2.80, p > .05, η2 =.14, for the target position. Contrary to our prediction, the mood
induction x target position was also nonsignificant, F < 1. The results are visualized in Figure
4.
Statistical error rate analysis
The total number of errors in response to the target was 7.31%. The errors were computed for
each condition for each participant. A repeated-measures ANOVA was conducted on these
totals with target location (top vs. bottom) as the within subject factor and mood induction
condition as the between subject factor. The factor target location was not significant, F(1, 18)
= 2.07, p > .05. There was no significant main effect of mood induction condition, F < 1, and
the interaction between target location and mood induction condition was also nonsignificant,
F(1, 18) = 1.12, p > .05. The results mirrored hose found in the analysis of the reaction time
data.