Odor-Induced Saltiness Enhancement: Insights Into The Brain Chronometry Of Flavor Perception

Highlights

• An odor can enhance a taste when the brain integrates both perceptions into a flavor percept.

• Participants were stimulated in-mouth with a close-to-real soup model, with reduced salt and a beef stock odor.

• Late cognitive P3 peak (600 ms) of ERP was delayed when odor-induced saltiness enhancement occurred.

• Early perceptual P1 peak (100 ms) did not vary in amplitude or latency during odor-induced saltiness enhancement.

• Odor-induced saltiness enhancement likely occurs in high-level integration areas and not in primary sensory cortices.

Abstract

Flavor perception results from the integration of at least odor and taste. Evidence for such integration is that odors can have taste properties (odor-induced taste). Most brain areas involved in flavor perception are high-level areas; however, primary gustatory and olfactory areas also show activations in response to a combination of odor and taste. While the regions involved in flavor perception are now quite well identified, the network's organization is not yet understood. Using a close to real salty soup model with electroencephalography brain recording, we evaluated whether odor-induced saltiness enhancement would result in differences of amplitude and/or latency in late cognitive P3 peak mostly and/or in P1 early sensory peak. Three target solutions were created from the same base of green-pea soup: i) with a “usual” salt concentration (PPS2), ii) with “reduced” salt (PPS1: −50%), and iii) with reduced salt and a “beef stock” odor (PPS1B). Sensory data showed that the beef odor produced saltiness enhancement in PPS1B in comparison to PPS1. As the main EEG result, the late cognitive P3 peak was delayed by 25 ms in the odor-added solution PPS1B compared to PPS1. The odor alone did not explain this peak amplitude and higher latency in the P3 peak. These results support the classical view that high-level integratory areas process odor–taste interactions with potential top-down effects on primary sensory regions.

Introduction

We experience food as a unitary perception, which we commonly call “taste”. This common “taste” is actually a holistic perception of at least olfactory and gustatory inputs, called “flavor perception”. Odor-induced taste enhancement (OITE) is a phenomenon that derives from the integration of taste and odor into flavor perception. For example, it was shown that a strawberry odor could increase the sweetness of a whipped-cream with sucrose. This result was first highlighted by Frank and Byram (1988). They also defined a fundamental principle of OITE, namely that only congruent odors and tastes would produce OITE, therefore pointing at the role of experience in shaping OITE. Indeed congruent, familiar, and complex flavor mixtures -which are more prone to be perceived as configural units- are more effective in producing OITE (Prescott et al., 2004, Small and Prescott, 2005, Labbe et al., 2006). Several independent labs have later replicated this finding and further demonstrated odor-induced taste enhancement of other tastes (Frank and Byram, 1988, Schifferstein and Verlegh, 1996, Sakai et al., 2001, Djordjevic et al., 2004, Prescott et al., 2004, Lawrence et al., 2009, Wang et al., 2019). OITE is, therefore, a reliable phenomenon. Other odor–taste interactions have also been established, such as the taste-induced odor enhancement (i.e., the reverse effect of OITE) (Lim et al., 2014, Linscott and Lim, 2016). In our study and the discussion of the results, we focused on the odor-induced saltiness enhancement only.

Most studies on OITE used water with sugar or salt and aroma, which produced non-ecologically relevant, unfamiliar and likely unpleasant perceptions (Prescott et al., 2004, Welge-Lüssen et al., 2005, Marshall et al., 2006, Prescott and Murphy, 2009, Welge-Lussen et al., 2009, Lim and Johnson, 2011, Lim and Johnson, 2012, Seo et al., 2013). To overcome this issue, one can use close-to-real food models, which produce more familiar and holistic food representations. It may also facilitate the OITE with appropriate congruent aroma and smooth out significant hedonic variations that could mask subtle integration mechanisms (Prescott, 1999, Small, 2012, Mroczko-Wąsowicz, 2016, Thomas-Danguin et al., 2016). Other studies used more complex and familiar food models. For example, ethyl 2-methyl butanoate was used in a fruit juice to enhance sweet perception (Barba et al., 2018). In another study, the authors used a sardine aroma to significantly enhance saltiness in a cheese model (Syarifuddin et al., 2016). In the present study, we studied odor-induced saltiness enhancement (OISE). A salty food model has been designed from a green-pea soup base, which was chosen for its composition with a negligible quantity of salt and an easily identifiable odor component. Five conditions with different salt and aroma quantities were selected to produce OISE according to previous results (Lawrence et al., 2009, Nasri et al., 2013). The first condition was the soup added with a standard (usual) level of salt (6.25 g/L), to record the most familiar level of saltiness in this kind of product and to test whether OISE could reinforce saltiness up to a “normal” saltiness intensity. The second solution was 50% salt reduced. The third condition, which is the target beverage, was reduced in salt (50%) and supplemented with a “beef stock” odor chosen for its potential to increase saltiness perception. Finally, two controls were tested, the base soup alone and the base soup with the odor component, to test the effect of added odor in the food model.

Endogenous mechanisms produce OITE in the brain. Several functional magnetic resonance imaging (fMRI) studies have investigated brain areas involved in flavor perception, leading to the identification of a relatively consensual flavor network (Rolls and Baylis, 1994, Rolls, 1997, de Araujo et al., 2003, Small and Prescott, 2005, Seo et al., 2013, Seubert et al., 2015). In these studies, supra-additive activations for the odorant-tastant mixture were found in high-level areas, in the orbitofrontal cortex, the dorsal mid-insula, and the perirhinal cortex (de Araujo et al., 2003, Seo et al., 2013, Small et al., 2013, Seubert et al., 2015). However, odor–taste convergence was also found in the primary gustatory cortex, more precisely in the anterior insula and frontal operculum (de Araujo et al., 2003, Seubert et al., 2015). Regarding these fMRI results, two views exist i) one consists in a hierarchical integration, starting with a parallel unimodal encoding of odor and taste in their respective cortices and further elaborated by higher-order unimodal zones, before converging onto multisensory integrative areas to form the flavor perception; ii) while the second proposed that odor and taste are already integrated into primary olfactory and gustatory cortices (Small and Prescott, 2005, Verhagen and Engelen, 2006, Verhagen, 2007, Prescott, 2012, Small et al., 2013).

To understand whether odor and taste already interact in the primary cortices or later in higher cortices, we need to study the chronometry of odor–taste integration and interaction. Electroencephalography (EEG) is of particular interest to gain insights into these questions. It permits quantitative measures of global brain activations with a resolution of milliseconds. Olfactory-gustatory event-related potentials (ERPs) give access to the chronometry of interactions between gustatory and olfactory cortices. To do so, one should select appropriate stimuli, which permit to isolate variables of interest (e.g., real saltiness or induced saltiness). An event-related potential (ERP) is a sequence of brain components identified by the maximum amplitudes of a series of positive and negative peaks, from P1 the earliest to P3 the latest measurable. ERP reflects the different steps of information processing in the cortex. The peak amplitude and latency provide a quantitative measure of the intensity and/or amount of neurons discharging in a synchronized way in response to the stimulus provided at t0. The early P1 peak mainly occurs in primary sensory areas and represents the processing of sensory and chemically related properties of food. The late P3 peak occurs mostly in high integratory and cognitive areas and illustrates endogenous processing such as emotional, semantic, decisional, and top-down mechanisms towards primary regions.

While an extensive literature exists on food-related visual event-related potentials (ERPs) (for review, see Carbine et al., 2018), very few studies were based on the senses directly involved in flavor perception: olfaction and gustation. To the best of our knowledge, no EEG studies showed the brain mechanisms of odor-induced taste enhancement. However, Welge-Lüssen et al., 2005, Welge-Lüssen et al., 2009 designed two studies to show the effect of taste (sucrose or lemon pulp) on odor (vanilla) or trigeminal perception (elicited by CO2) respectively. In these studies, participants were sucking on a taste dispenser when an odor was sent orthonasally (Welge-Lüssen et al., 2005) or retronasally (Welge-Lüssen et al., 2009) with an olfactometer. This moment corresponded to the start of the ERP, which therefore highlighted the odor processing modulated by the taste. Although sensory results did not show odor or taste enhancement, ERPs tended to higher amplitude and reduced early and late peak latencies (P1 and P3), only when the taste matched the odor. These two peaks represent the earliest and latest observable brain mechanisms of the evoked potential measured with EEG. Therefore, Welge-Lüssen's studies showed that taste sped up the processing of a congruent odor from the very first processing mechanisms (P1 peak). Although these results did not permit observing supra-additive effects for a flavor mixture compared to its odorant-tastant components, they were interestingly discussed in terms of priming. To observe supra-additive effects, one should synchronously stimulate the olfactory and gustatory cortices and compare activation for the mixture to the single components. Recent electrophysiological results in animal reconsidered the classical view of late odor–taste integration. They indeed showed activations in a region of the primary olfactory cortex, i.e., the piriform cortex, in response to sucrose (sweet taste), which the authors considered early interactions (Maier et al., 2012, Maier et al., 2015, Maier, 2017). These results, therefore, challenge the classical theory of late brain interactions between odor and taste and suggest that primary olfactory and gustatory areas may interact as early as the primary EEG peaks such as P1 (100–200 ms).

Therefore, we addressed the chronometry of the integration of odor and taste into flavor perception, in humans, by studying the chronometry of brain mechanisms leading to OISE. The classical view, which consisted of a hierarchical integration of flavor, from primary gustatory and olfactory areas to secondary or tertiary integratory cortices, has been further expanded to explain OITE (Verhagen and Engelen, 2006, Verhagen, 2007, Small, 2008, Prescott, 2012). Following the integration of odor and taste into the flavor, top-down feedback may control activations in gustatory areas producing an increased endogenous perception of saltiness intensity. Following this reasoning, the odor-induced saltiness enhancement should be observed only on the ERP's late components. Therefore, we hypothesized that differences of amplitude and/or latency could be observed mostly on the latest peak of olfactory-gustatory ERP (the late P3 peak) and not on the P1 peak involving brain circuits responding to exogenous properties of the food such as tastant concentration. In the study, we did not address whether retronasal odor stimulation is necessary for the supra-additivity of the flavor solution. Still, to avoid any potential bias, we used only retronasal odor perception. Because of the presumed importance of oral referral in flavor perception (Small, 2008, Spence, 2016), which is supported by EEG results (Welge-Lüssen et al., 2009, Masaoka et al., 2010), participants should be stimulated in-mouth so that aromas could be perceived through the retronasal route. However, the aroma perception is supposed to be maximal when participants are swallowing. A dedicated paradigm was designed to account for the swallowing artifacts and the need for synchrony between odor and taste perceptions.