Thirty-two of the reviewed articles studied aspects of olfaction; 22 employed neuroimaging techniques, eight used behavioural methods and two performed meta-analyses. fMRI was the most used neuroimaging method, with ten neuroimaging papers using a task-based fMRI method and seven using resting-state fMRI. The bias for employment of fMRI methodology, which has exceptional spatial resolution, reflects the common research theme of localising olfactory processes within the brain. As the primary and secondary olfactory regions have been extensively documented prior to 2002 (see Figures 4, 5), many of the reviewed papers instead seek to localise specific higher level cognitive olfactive processes. The regions associated with different olfactory-related cognitive processes are summarised in Table 2 and Figure 6.

Another common theme involved localisation of olfactory function and functional changes within specific populations. Eight studies evaluated participants with olfactory dysfunction, and other studied population groups included student vs. expert perfumers, younger and older participants, early and late blind participants, and participants with temporal lobe epilepsy.

With olfactory regions well established, another common approach was to characterise the involvement of these regions within the wider network. The increasing popularity of functional network analyses within cognitive neuroscience is allowing characterisation of localised brain region function in cognitive processes which are lost in subtractive models. Ten of the reviewed studies employed a functional connectivity or network-based analysis approach, and it is likely that olfactory research will continue to see an increase in this approach, as is seen in other cognitive neuroscience domains.

A prominent research theme across the reviewed articles was the study of hedonics; ten studies considered the impact of odour valence on olfactory processing. Seven of these articles directly contrasted odour valence: two behavioural (Hudry et al., 2014; Kärnekull et al., 2021) and five neuroimaging (Royet et al., 2003; Bensafi et al., 2007; Callara et al., 2021; Gorodisky et al., 2021; Torske et al., 2022). A further three papers involved discussion of the impact of odour hedonicity on olfactory processes, but did not directly manipulate odour pleasantness (Bensafi and Rouby, 2007; Plailly et al., 2007; Morrot et al., 2012). Royet et al. (2003) further extended this by exploring the impact of handedness, gender and active hedonic judgements on hedonic odour processing. The lateral aspect of the left OFC was implicated in mediating the conscious assessment of odour pleasantness, with this lateralisation being particularly pronounced in female participants. All seven of the neuroimaging papers which included themes of hedonic odour processing also cited OFC involvement. Other regions commonly cited for their involvement in hedonic odour processing included the piriform cortex, cingulate gyrus (CgG), superior temporal gyrus (STG), amygdala and insula.

Royet et al. (2003) identification of differential involvement of the left and right OFC in olfactory processing also supports evidence of the lateralisation of olfactory processing. First proposed by Broman et al. (2001), the differential involvement of the left and right hemispheres in olfactory processing remains a pertinent topic of discussion within olfactory research. Broman et al. suggested the right hemisphere is involved in low-level perceptually based odour processing and encoding, and the left hemisphere is associated with higher-level cognitive-based odour recognition processes and semantic interpretation. Royet et al. findings support this theory, with the left OFC expressing greater involvement that the right OFC in the judgement of odour pleasantness. Nine other reviewed articles discussed or presented evidence to support this lateralisation in olfactory processing (Royet et al., 2003; Djordjevic et al., 2005; Bensafi et al., 2007; Plailly et al., 2007; Hudry et al., 2014; Kollndorfer et al., 2015b; Zhou et al., 2019; Douaud et al., 2022; Kretzmer and Mennemeier, 2022; Eek et al., 2023).

Hudry et al. (2014) provided further evidence of left hemisphere dominance in semantic and emotional olfactory processing by studying participants with unilateral temporal lobe epilepsy (TLE). Their findings highlighted the privileged role of the left hemisphere for emotional and semantic processing, with left TLE participants judging odours as less pleasant and exhibiting greater difficulty with identification. Furthermore, the reported advantage for judging odour familiarity during right nostril stimulation validates the role of the right hemisphere in encoding the sensory percept of an odour; familiarity ratings largely reflect the clarity of perceptual processing (Broman et al., 2001; Royet, 2004).

Kollndorfer et al., 2015b evaluated three functional networks involved in olfactory processing labelled as the olfactory network, the somatosensory network, and the integrative network. They reported the olfactory network was relatively symmetrical across both hemispheres, whereas the somatosensory network expressed significantly greater right hemisphere recruitment and the integrative expressed a clear left hemisphere bias. Zhou et al. (2019) performed a laterality index analysis to quantify functional asymmetry of the olfactory processing. Similarly to Kollndorfer et al., they did not identify any significant asymmetry across the primary olfactory network. These findings, suggest that odours are perceived equally by both hemispheres, but that each hemisphere proceeds to encode different aspects of the odour: the right hemisphere encoding the olfactory perceptual experience, and the left hemisphere encoding emotional and semantic interpretations of the odour.

The most notable theme within the reviewed olfactory research was the characterisation of olfactory memory. Fourteen of the reviewed papers weighed in on the debate regarding olfactory memory processes. Whilst only three of the reviewed papers actively studied olfactory memory (Plailly et al., 2007; Meunier et al., 2014; Eek et al., 2023), eleven papers were able to apply their findings to contribute further knowledge to the discussion of olfactory memory (Djordjevic et al., 2004; Bensafi and Rouby, 2007; Plailly et al., 2011; Hudry et al., 2014; Kollndorfer et al., 2015a; Zhou et al., 2019; Callara et al., 2021; Infortuna et al., 2022; Torske et al., 2022; Muccioli et al., 2023; Perszyk et al., 2023). Memory of odours and olfactory experiences presents a unique case of memory encoding and recall compared to other sensory modalities (Stevenson and Case, 2005; White, 2009; Eek et al., 2023). As such, it is understandable that there is great interest in the study of olfactory memory, and it remains such a prominent topic of research within olfactory research.

Olfactory memories are highly resistant to forgetting over time and often experienced with higher emotional intensity (Roediger et al., 2017). As shown by Callara et al. (2021) and Zhou et al. (2021), the olfactory system has connections with both the amygdala and hippocampus. Whilst other sensory systems must relay through the thalamus (Eek et al., 2023), the olfactory system is directly communicating with centres associated with emotion and memory. Callara et al. (2021) identified the OFC as the node with the highest inflow during olfactory stimulation, noting its key role in olfactory perception. They also identified strong interactions between the OFC and brain regions associated with emotion and memory. They conclude that these connections may be responsible for the enhanced encoding and emotional intensity of olfactory memory.

Discourse surrounding odour memory is inherently associated with the lateralisation of olfactory processing. The same left-right dichotomy in odour processing appears to be mirrored within odour memory encoding and recall (Broman et al., 2001; Royet, 2004; Hudry et al., 2014). The “dual process theory” describes two memory processes contributing to stimulus recognition: “familiarity,” described as perceptual recognition of an odour related to implicit or unconscious memory, and “recollection,” described as conceptually driven recognition along with contextual information retrieval involving explicit or conscious memory (Royet, 2004; Eek et al., 2023). These memory processes are associated with the right and left hemispheres respectively, mirroring the described laterality of olfactory processing (Royet, 2004; Hudry et al., 2014). Hudry et al. (2014) study particularly highlights the complex interplay between the hemispheres in the recollection and familiarity of odours; odour identification was impaired in participants with left TLE, whereas odour familiarity ratings were associated with a clear right-nostril advantage.

Another point of discussion within olfactory memory research pertains to the existence of an olfactory working memory capacity (White, 1998; Stevenson and Case, 2005). The reviewed literature presents a general consensus to support the existence of a working memory. One method to interrogate olfactory working-memory is through odour discrimination; discrimination between successive odour stimuli requires working memory involvement to hold the perceptual trace of the first stimuli for comparison with the subsequent odour presentation. Plailly et al. (2007) employed an odour discrimination paradigm inspired by the n-back task, a common paradigm used to investigate working memory, to evaluate olfactory working-memory. Authors identified activations in the left IFG and OFC associated with the maintenance of the first odour perceptual trace, demonstrating the existence of an olfactory working-memory capacity.

Working-memory capacity in the olfactory domain can also be investigated via tasks which require the maintenance of a neural representation of olfactory stimuli. For example, Hucke et al. (2023) used a combined EEG and fNIRS methodology to investigate the requirement of trigeminal stimulation for neural representation of odour source localisation. The involvement of somatosensory cortices during localisation of odour stimuli indicates the dorsal network involvement in processing where a stimulus occurs, as has been extensively documented during visual processing, also extends to olfactory processing (Frasnelli et al., 2012; Hucke et al., 2023). These results also support the sensorimotor recruitment models of working memory whereby the systems involved in the sensory perception of stimuli can also hold a short-term representation of sensory information (D’Esposito and Postle, 2015). This provides further support for the existence of an olfactory working-memory capacity which mirrors that of other sensory modalities.

The prominence of memory discussion in olfactory processing also appears to be closely related to the research theme of odour hedonics, as hedonic judgements are mainly driven by memory and semantic smell identification (Schleidt et al., 1988; Sucker et al., 2007). The performance of hedonic odour judgement, particularly of unpleasant odours was consistently associated with left hemisphere involvement within the reviewed literature. Given the evidence surrounding the lateralisation of odour memory, it appears that this left hemisphere bias is indicative of sematic and contextually driven odour recollection processes, mediated by the left hemisphere.

All the reviewed neuroimaging literature reported activation within at least one of the documented olfactory processing regions. The reviewed literature presents a consensus as to the lateralisation of olfactory function, with the right hemisphere associated with low-level olfactory perceptual processing, and the left hemisphere associated with higher level cognitive olfactory processing including hedonic judgements, odour naming, semantic interpretation and olfactory memory. The three most prominent themes within the reviewed literature; hedonic odour perception, lateralisation of odour processing and olfactory memory, all appear to be very closely related. Hedonic odour perception was associated with mostly left-lateralised regions including left OFC, CgG, STG, piriform, and amygdala, and bilateral insulae. Multiple studies reported activation in regions associated with memory recall and working memory, including the PrC, SPL and IFG. The reviewed studies appear to provide support for the existence of an olfactory-specific working-memory capacity. This supports the notion that olfactory imagery is mediated by the same mechanisms underlying other imagery modalities, and hence is a “true” form of sensory imagery.

Twenty-two articles studied olfactive imagery. Eleven papers employed neuroimaging methods, ten employed behavioural methods and one performed a meta-analytic review. Nine neuroimaging papers used task-based fMRI, one used PET and one used TMS and EMG. Once again, the dominance of fMRI in olfactory imagery research appears to reflect a common aim of localising olfactory imagery regions within the brain. A summary of brain regions associated with olfactory imagery is presented in Table 3 and Figure 7.

Five neuroimaging papers sought to localise regions associated with olfactory imagery by contrasting imagery and perception (Djordjevic et al., 2005; Bensafi et al., 2007; Plailly et al., 2011; Leclerc et al., 2019; Perszyk et al., 2023). Leclerc et al. (2019) identified a largely left lateralised network including left DLPFC, IFG, IPS, angular gyrus and pre-SMA, and right frontal pole and IFG which was more active during odour imagery than during odour perception. This appears to mirror the prominent discourse around the lateralisation of olfactory processes within olfaction research; the findings of a mostly left-lateralised network associated with olfactory imagery further corroborates the left hemisphere is involvement in the higher level cognitive olfactory processes including odour memory and semantic labelling. Somewhat conversely, Djordjevic et al. (2005) compared olfactory perception and olfactory imagery, finding odour imagery efficiency scores were significantly correlated with rCBF increases in right anterior and posterior OFC. The authors concluded that this positive correlation suggests successful odour imagery occurs when the brain treats odour images the same as perceived odours.

Another approach to localise olfactory imagery regions contrasted olfactive imagery with imagery in other sensory modalities such as visual or auditory imagery. Four papers contrasted sensory specific imagery with modality general imagery regions and identified the left lateralised imagery network is modality general, but that there are also regions associated with olfactory-specific imagery (McNorgan, 2012; Flohr et al., 2014; Leclerc et al., 2019; Han et al., 2022). McNorgan (2012) performed a meta-analysis of articles studying uni- and multi-sensory imagery to localise imagery general and modality specific brain regions. They analysed 65 research reports across olfaction, audition, gustatory, motor, tactile, visual-colour, visual-form and visual-motion. Analysis identified a general imagery network of eight, mostly left lateralised regions. Four left-lateralised clusters exclusively associated with olfaction were identified in the anterior cingulate, hippocampus, amygdala and SPL. Similarly, Han et al. (2022) identified olfactory imagery was associated with greater activation in bilateral PrC (SPL) and superior occipital cortices, left hippocampus and right SFG than visual imagery. The authors concluded this increased involvement of the PrC, superior occipital regions (cuneus) and hippocampus in the odour imagery condition suggest that odour imagery may rely more on memory retrieval processes than visual imagery (Figure 8).

Most of the reviewed papers evaluated olfactory imagery in a healthy, non-clinical population. Due to the wide heterogeneity in olfactory imagery abilities across the general population, it is understandable that many reviewed studies still seek to characterise olfactory imagery in the general population, rather than focusing on subgroups with potentially atypical olfactory imagery. However, six papers did investigate specific populations. Studies which included specific populations included dysosmic or anosmic participants (Flohr et al., 2014; Kollndorfer et al., 2015a; Tomasino et al., 2022), young vs adult participants (Arshamian et al., 2020) and student vs expert perfumers (Plailly et al., 2011; Royet et al., 2013a,b). A common research theme within these papers was to consider the impact of olfactory exposure on olfactory imagery abilities. All the reviewed papers agreed that varying expertise was associated with the recruitment of different brain regions for olfactory imagination.

One potential reason for this difference in regions may be due to differences in imagery generation techniques. Han et al. (2022) unique paradigm investigated differences odour imagery generation across varying olfactory imagery abilities by employing short vs. long odour imagery generation times. Participants with lower olfactory imagery expressed stronger activation in the left SMA and right SFG in the short olfactory imagery condition than in the long olfactory imagery condition, brain regions involved in modality general mental imagery (McNorgan, 2012; Zvyagintsev et al., 2013). This increased activation of multisensory regions may indicate participants with lower olfactory imagery abilities formed mental images including other sensory modalities to facilitate olfactory imagination.

Another possible reason for these differences may be differences in retrieval effort of olfactory memories. Plailly et al. (2011) identified a bilateral network of regions including the right MFG which expressed reduced imagery-induced activation with expertise. They concluded that the activation decrease associated with increased olfactory imagery performance are reflective of the “retrieval effort” (Tulving, 1985); student perfumers at the beginning of their career must deploy a greater level of processing resources to retrieve the olfactive image than expert perfumers. These findings are further extended by Flohr et al. (2014) who identified increasing activation in bilateral DLPFC (MFG) associated with olfactory loss, and that the degree of DLPFC activation varies with longevity of olfactory dysfunction. Flohr et al. hypothesise this varying recruitment of DLPFC is the result of greater recruitment of working memory resources based on similar observations amongst the visually impaired (Dulin et al., 2011), also providing further evidence of an olfactory working memory capacity, and supporting the conclusion of Plailly et al. (2011) that increasing activation within these regions is indicative of greater retrieval effort correlating with lower olfactory expertise.

Royet et al. (2013a,b) re-analysed Plailly et al. (2011) data to identify changes to functional coactivation between 22 ROIs identified by Plailly et al., hypothesising that increasing olfactory expertise would be associated with greater connection across olfactory memory regions. They identified professional perfumers demonstrated significantly greater coactivations between MFG and the rest of the olfactory imagery network, and significantly lower coactivation between the PrC and rest of the imagery network than student perfumers. They concluded these changes to connectivity reflect differences in the recall mechanisms underlying olfactory imagery between student and expert perfumers. According to “multiple trace theory” of memory consolidation, retrieval-related activation of the hippocampus reduces over time, with more involvement of prefrontal cortex regions in the recall of more mature memories; retrieval of some distant memories can have no hippocampal involvement. The increased connectivity of the middle frontal gyrus with olfactory and memory regions in the expert group is likely indicative of post hippocampal memory recall in the expert group. In contrast, the increased coactivation of the PrC with memory and olfactory regions in the student group is indicative of allocation of top-down attentional resources to memory retrieval. Involvement of the superior parietal lobe, including the PrC, during memory recall has also been associated with lower confidence in the accuracy of mental imagery (Ciaramelli et al., 2008).

A key finding across the reviewed literature is the involvement of key memory retrieval and working memory regions in olfactory imagery. This supports the argument that olfactory imagery is a true form of sensory imagery. The involvement of memory regions is also demonstrated to vary with varying levels of olfactory expertise. It is hypothesised that this reflects differences in the mechanisms of retrieval, and use of polymodal imagery to facilitate odour imagery generation. Localisation of olfactory imagery regions compared to olfactory perception reveals a largely left lateralised olfactory imagery network including left DLPFC, IFG, IPS, angular gyrus and pre-SMA. This reflects the lateralisation of olfactory function as proposed by Broman et al. (2001) and prominently discussed within olfactory research that the right hemisphere is associated with the sensory perception of odours, and the left hemisphere is involved in the higher level cognitive olfactory processes including odour memory and semantic labelling. However, many of the regions identified in this network appear to be modality-general imagery regions. When contrasted with imagery in other sensory modalities, olfactory specific activity is observed in the anterior cingulate, hippocampus, amygdala and SPL, regions which have also been implicated in memory recall. It is likely that the enhanced involvement of memory recall regions within olfactory imagery when compared to other modalities is the result of greater retrieval effort required to form an olfactory mental image, and top-down attention direction towards the intended modality within an involuntary polymodal mental image formed to facilitate olfactory imagery generation.

Thirteen studies investigated crossmodal interactions between vision and olfaction. Twelve of these papers employed neuroimaging techniques and one used only behavioural measures. The most common imaging modality was fMRI, employed by seven of the reviewed papers. This is likely reflective of a strong research aim of characterising the regions involved in crossmodal interactions. One study used fNIRS to investigate crossmodal colour-odour correspondances. The use of fNIRS in this study allowed the investigation of colour-odour correspondances using a unique paradigm which has not been used in previous neuroimaging investigation of crossmodal visual-odour correspondances. Regions associated with crossmodal interactions are summarised in Table 4 and Figure 9.

The most commonly used paradigm involved presenting participants with unimodal vs bimodal visual and olfactory stimuli. This protocol was used by six studies, five fMRI studies (Gottfried and Dolan, 2003; Österbauer et al., 2005; Ripp et al., 2018; Stickel et al., 2019; Schicker et al., 2022) and one behavioural study (Amsellem et al., 2018). Within the bimodal condition, all of these studies presented the bimodal stimuli as congruent or incongruent pairs. Additionally, Amsellem et al. (2018) included semi-congruent and semi-incongruent conditions. They achieved this by selecting two target odours and creating three additional blended fragrances with varying ratios of the two target odours. Through this, they were able to demonstrate that congruence between visual and olfactory stimuli is not a dichotomy, but rather that participants were able to detect the nuances of varying degrees of congruence, which impacted upon pleasantness ratings.

In addition to unimodal vs bimodal conditions, three papers (Gottfried and Dolan, 2003; Ripp et al., 2018; Stickel et al., 2019) also included pleasant and unpleasant valence conditions. This resulted in four bimodal, four unimodal and one baseline condition. From this Gottfried and Dolan (2003) contrasted these conditions to identify brain regions associated with olfaction, pleasant and unpleasant odour perception, olfactory-visual interactions and congruence of olfactory-visual stimuli. Both Ripp et al. (2018) and Stickel et al. (2019) performed connectivity analyses. Stickel et al. (2019) investigated multisensory integration using DCM to analyse information exchange during bimodal olfactory-visual or auditory-visual stimulation. Using three key regions identified from the unimodal visual (cuneus), unimodal olfactory (amygdala) and bimodal congruent (IPS) conditions, Stickel et al. modelled the network linked to integration of visual and olfactory stimuli. Their model composed of a driving input of bimodal olfactory-visual stimulation to the IPS and nonlinear modulations from IPS to the reciprocal cuneus ↔ amygdala connection (Figure 10). Their results showed an overlapping network of brain regions involved in multisensory integration of olfactory-visual and audio-visual information. They also demonstrate the IPS modulates changes between areas relevant to sensory multisensory perception by exerting top-down control over primary sensory regions.

Ripp et al. (2018) also used a unimodal vs bimodal paradigm with congruent and incongruent, and pleasant and unpleasant conditions, and a graph theoretical network based functional connectivity analysis. Ripp et al. (2018) identified six nodes which expressed significantly stronger functional connectivity in the bimodal condition than the combination of unimodal conditions. Bimodal presentation of odour and pictures, collapsed across valence, was associated with significantly greater functional connectivity between the right putamen ↔ right insula, PrC ↔ left SMG and left MOG ↔ left IFG. Involvement of the right insula and putamen has been observed in previous studies of multisensory integration, regardless of sensory modality (Banati, 2000; Bushara et al., 2001; Olson et al., 2002; Naghavi et al., 2007; Renier et al., 2009), leading the authors to conclude that this connectivity between the insula and right putamen is part of a functional multisensory integration specific network. Involvement of the PrC, as cited in Royet et al. (2013a,b), is likely indicative of top-down facilitation of memory retrieval. Ripp et al. proposed that the increased connectivity between the PrC and SMG is indicative of memory retrieval and maintenance of the retrieved memory within a phonological store, once again supporting the evidence for an olfactory working memory. The left IFG has also been shown to be associated with odour working memory, semantic interpretation and odour naming (Djordjevic et al., 2005; Plailly et al., 2007). Ripp et al. suggested that the increased connectivity of the left MOG, a visual processing region, and left IFG allows the matching of visual information with odour semantic information. They further propose that retrieved odour memory information, held in the phonological store, along with visual information and semantic information from the left MOG and IFG are passed via the inferior fronto-occipital fasciculus, a large white matter tract connecting the frontal, temporal and occipital lobes, to the temporal association cortex where this information is fused into a multisensory percept.

Four fMRI studies involved connectivity-based analyses, with two employing Dynamic Causal Modelling (DCM) (Novak et al., 2015; Stickel et al., 2019), one employing psychophysical interaction analysis (Sijben et al., 2018) and one performing graph theoretical network analysis (Ripp et al., 2018). As with olfaction, this reflects a trend towards investigation of network-based interactions underlying multisensory integration. As multisensory integration and crossmodal correspondences require the integration of information from multiple networks, including sensory specific processing networks and memory networks, Sijben et al. (2018) argue that connectivity-based analyses are a better tool to characterise these processes than subtractive analysis models. Similar to Amsellem et al. (2018) and Sijben et al. (2018) included a semi-congruent condition to investigate the impact of semantic congruence on olfactory-visual integration. Their results indicated a differential connectivity of parcellations of the IFG with seed regions from different networks involved in sensory and multisensory processing depending on the degree of congruence between the stimuli. This highlights the crucial role of the IFG in multisensory processing, potentially functioning as a hub for determining the degree of congruence between the stimuli. This supports Ripp et al. (2018) suggestion that IFG supplies odour working memory and semantic information for integration of visual-olfactory information. Increased connectivity with the putamen during congruent and semi-congruent multisensory processing also reflects previous findings of putamen involvement in multisensory integration. Using a connectivity approach, Sijben et al. (2018) were able to go beyond identifying regions involved with visual-olfactory integration, and instead were able to begin to describe the mechanisms of action within these regions.

Rather than using an image to provide visual stimulation, Österbauer et al. (2005) used colours. Odours and colours were presented in unimodal, bimodal congruent or bimodal incongruent form with participants responding as to how well the odour and colour “fit.” Unimodal odour presentation was associated with activity in primary and secondary olfactory regions: bilateral piriformis and amygdalae, putamen, right OFC and left insula. Using colour-odour congruency as an additional parametric modulator, Österbauer et al. identified a network of brain areas exhibiting increasing activity with higher perceived congruence. This network was entirely left lateralised and included OFC, IFG, gyrus rectus and anterior insula. Österbauer et al. also performed an additional contrast to identify regions which express superadditive responses to bimodal congruent stimuli, as described by Calvert et al. (2000, 2001), and Calvert (2001). Regions which express a greater response to bimodal stimulation than the addition of responses to unimodal stimulation conditions (i.e., olfactory-visual > olfaction + visual) are said to express linear superadditivity. Österbauer et al. also included behavioural ratings of colour-odour congruence within their superadditivity model such that BOLD response to colour-odour pairings which were rated as a “very good fit” were modelled as larger than the response to pairings which a “very bad fit.” They identified superadditivity effects of colour-odour stimulation within the SFG, ACC and OFC. Österbauer et al. observation of right OFC involvement in unimodal olfactory processing versus left OFC correlation with colour-odour congruence further supports Broman et al. (2001) theory that the right hemisphere is involved in low-level perceptually-based odour processing, and the left hemisphere associated with higher-level cognitive-based odour recognition and semantic interpretation.

Similarly, Yamashita et al. (2022) investigated colour-odour correspondences. Their use of fNIRS allowed investigation of colour-odour correspondences in a novel immersive paradigm; whilst previous research has typically employed presentation of small-field colour patches or display stimuli, using fNIRS allowed Yamashita et al. to sit participants within a booth illuminated in one of different colours. fNIRS monitoring was performed with two channels covering the left and the right OFC; the study compares the balance of left versus right OFC involvement in each of the stimulation conditions. Perception of pleasant and unpleasant fragrances, and crossmodal colour-odour stimulation were all associated with greater oxyhaemoglobin (HbO) change in the left OFC than right OFC, agreeing with previous findings of greater left hemisphere involvement in higher order cognitive olfactory processes (Broman et al., 2001; Hudry et al., 2014). Crossmodal presentation of odours resulted in greater OFC signal change than crossmodal presentation of odour names. As OFC has been shown to demonstrate superadditivity effects during crossmodal olfactory-visual stimulation (Österbauer et al., 2005), these results indicate that synaesthetically driven crossmodal correspondences are more harmonious than semantically driven correspondences. However, care must be taken when interpreting these results as indications of neural activity. HbO signals are more vulnerable to systemic artefacts which may artificially amplify signal changes and affect interpretation of results (Kirilina et al., 2012; Tachtsidis and Scholkmann, 2016; Dravida et al., 2018). To draw any firm conclusions, the study should be repeated with analysis of both oxy- and deoxyhaemoglobin (HbR) signals to ensure the signal changes are arising from neuronal activity rather than scalp level haemodynamics or other systemic artefacts.

The most common paradigm to investigate crossmodal interactions was a unimodal vs bimodal stimulation task. As expected, unimodal olfactory stimulation was associated with activity within primary and secondary olfactory regions. Bimodal olfactory and visual stimulation was associated with activity in a largely left lateralised network including OFC, IFG, gyrus rectus and insula. These findings further support the theory of lateralised olfactory processing proposed by Broman et al. (2001) that the right hemisphere mediates low level olfactory perceptual processes and the right hemisphere is mediating the higher level cognitive olfactory processes. The involvement of the left IFG reflects previous findings that the left IFG is associated with semantic recognition of odours and hedonic judgements. Superadditive effects were noted in the SFG, ACC and OFC. The involvement of the SFG and ACC provides further support for the role of memory networks within the creation of crossmodal percepts. Additionally, the differential involvement of right OFC in unimodal olfactory and left OFC in bimodal stimulation provides further support for the laterality of olfactory processing proposed by Broman et al. (2001). Functional connectivity analyses highlight the involvement of parietal regions including the IPS and PrC exerting a top-down control over primary sensory regions.