• brain connectivity
• dynamic causal modeling
• eeg
• hedonic olfaction
• sympathetic responses

Emotions can be elicited and perceived through different sensory processes. Although
research in the affective field has mainly focused on visual and auditory
channels, olfaction represents a powerful albeit underestimated channel to investigate
physiological and behavioral emotional responses.
In this PhD thesis, I have studied the brain cortical correlates of emotional olfactory
processing through the analysis of electroencephalographic signal (EEG). In particular,
since hedonic olfactory processing arises from the complex integration among hierarchically
organized networks, I focus on the application of EEG effective connectivity
techniques to investigate the modulatory effect of hedonic odors on the cortico-cortical
interactions. Particularly, the thesis builds on three experimental studies that have investigated
different aspects of the olfactory-driven emotional neural response through
the application of both data-driven techniques, such as Granger-Causality (GC)-derived
measures and physiologically-sound computational models such as Dynamic Causal
Modeling (DCM). In the first study, combining independent component analysis (ICA)
and clustering methods, we have inferred a group-common network of cortical nodes
involved in the hedonic olfactory task. Then, I have estimated their dynamic causal
interactions by combining the multivariate autoregressive (MVAR) modeling and the
application of GC-derived renormalized Partial Directed Coherence (rPDC). The insights
provided by this first study have been used as prior knowledge to build the architecture
of an olfactory cortical network upon which I have modeled the modulatory
effect of odors’ valence through DCM. In particular, I take advantage of the Parametric
Empirical Bayes (PEB) framework to infer the average modulatory effect of valence
on the group-connectivity, as well as significant interactions with the perceived level
of arousal. Moreover, I test for the hypothesis that well-known gender differences in
hedonic olfactory processing are reflected in the average group-connectivity. Finally, in
the third study, I have investigated an experimental scenario in which contextual hedonic
olfactory cues modulate the visual perception of ambiguous faces. In this scenario,
we have also investigated whether the presence or absence of a peripheral sympathetic
response elicited by the combined visual-olfactory stimulus can affect the neural processing
of the emotional stimulus. To this aim, I develop a novel methodological approach integrating sympathetic information provided by electrodermal activity (EDA)
into the framework of DCM and PEB.
Our results have provided several new insights into brain hedonic olfactory processing.
The application of both GC and DCM has shown that the processing of non-neutral
olfactory stimuli can result in different connectivity patterns, rather than differences
in cortical activation. Specifically, the administration of both pleasant and unpleasant
odors has produced significant interactions among cortical regions, whereas the neutral
stimulus has not affected connectivity. Furthermore, the application of DCM and PEB
has highlighted an interaction of arousal with the processing of unpleasant odors and
has corroborated the hypothesis that gender differences in hedonic odor processing are
also reflected by changes in brain connectivity. Finally, the novel EEG-EDA approach
has revealed a significant effect of sympathetic arousal and odors’ valence on specific
ERP components associated with the processing of faces. Particularly, the results have
shown interesting modulatory effects targeting the connectivity of the inferior temporal
gyrus (ITG) with other key regions already known to be involved in the processing of
faces and odors. These dynamics seem to suggest a role the ITG plays as a regulatory
region of sympathetic reflexes and physiological arousal.
This thesis contributes to a deeper understanding of the physiological processes underlying
hedonic olfaction. Moreover, it provides detailed methodological applications
of EEG effective connectivity to the scenario of olfactory stimuli. In particular, the
presented innovative EEG-EDA approach can be generalized to investigate the effects
of peripheral sympathetic responses on brain dynamics in any event-related scenario of
affective stimulation.