For many people, the most important parts of smell and flavor are the memories they evoke and the emotions associated with them. To illustrate the impact these have, we can do no better than start with Marcel Proust. An understanding of the brain mechanisms involved throws new light on his classic story:
But when from a long-distant past nothing subsists, after the people are dead, after the things are broken and scattered, taste and smell alone, more fragile but more enduring, more unsubstantial, more persistent, more faithful, remain poised a long time, like souls, remembering, waiting, hoping, amid the ruins of all the rest; and bear unflinchingly, in the tiny and almost impalpable drop of their essence, the vast structure of recollection.
The quotation comes from Swann’s Way, the first volume of his great novel Remembrance of Things Past. In it, Proust describes how from the aroma of a tea-soaked biscuit called a madeleine a powerful memory of his childhood flooded back to him. This has become a cliché for the memory that springs suddenly and purely into mind after a long period of forgetting. However, several years ago a young English professor of my acquaintance carried out a close textual analysis that revealed that the memory was not sudden, but was the result of considerable effort as described by the author himself. That made me start to wonder what was happening during this time in Proust’s brain, between the initial sensing of the taste of the tea-soaked madeleine biscuit and the memory that was finally called forth.
An understanding of the neural pathways in the sense of smell and the perception of flavor starts to provide an answer. Several years ago, Kirsten Shepherd-Barr and I combined the textual analysis and the neural pathways to reinterpret this famous passage, and I update that reinterpretation here. If you have followed the argument in preceding chapters, you will recognize the principles as they unfold; if you have plunged in here, you will get the whole story as it applies to this episode.
Activating Proust’s Brain
We start with the knowledge that the taste of a madeleine must be mostly due to its smell. The stimulus for Proust’s taste experience was therefore primarily the odors emanating from the mixture of pastry crumbs soaked in tilleul, the aromatic lime-scented infusion made from linden blossoms. What, then, might these smells have been?
It is sometimes argued that Proust dithered over exactly what kind of biscuit it was that stimulated his reverie, but it really does not matter. A traditionally made madeleine, in addition to possessing odor molecules that arise from the butter and eggs, would include several types of “aroma essences.” Flavors of foods are enhanced by heating and dissolving in water, which increase the vapor pressure so that volatile molecules are released into the air or within the mouth. Thus, as children learn, a humble pastry gives off its aromas with greater effect when its crumbs are dissolved in hot liquid. The aromas in a madeleine would include vanilla and several types of related odor molecules in the lemon, such as citral and limonene, which belong to the terpenes, a family of essential oils secreted by plants. As described in chapter 4, they are highly volatile, consisting of 5 carbon atom units linked together in various shapes and with various reactive functional groups such as esters, alcohols, and acids. The other source of olfactory stimulation in Proust’s concoction was the tilleul, which contributed its own scent.
How do these molecules give rise to a smell perception? As we have seen, receptor molecules in the fine hairlike cilia lying in the mucus are stimulated by the smell molecules as they are inhaled into the nose (the orthonasal route), and also by the smell molecules released from within the mouth that rise into the nasal cavity from the back of the mouth (the retronasal route). It is by this latter route, after the narrator has taken the brew into his mouth, that the smell molecules are released and carried by the warm and humid air of his nasopharynx to his olfactory sensory cells.
The smell molecules, absorbed into the mucus, act on receptor molecules in the cilia membranes. These in turn initiate the cascade of microkicks from one signaling molecule to the next to change a membrane protein formed around a tiny channel that lets electric charges f ow through it. This alters the electrical potential across the cell membrane, leading to the discharge of impulses in the cell that is conveyed through its long fiber (axon) to the first relay station in the brain.
The narrator’s mouthful of crumb-laden tea thus activates a range of receptors tuned to the different volatile components, leading to impulse discharges that carry the information to the brain. But in addition to activating impulses, the signaling cascade in the receptor cells also contains a number of pathways for controlling the sensitivity of the sensory response. Repeated stimulation brings about desensitization of a second messenger pathway. In Swann’s Way, Proust appears to be describing precisely this effect:
I drink a second mouthful, in which I find nothing more than in the first, then a third, which gives me rather less than the second. It is time to stop; the potion is losing its magic. It is plain that the truth I am seeking lies not in the cup but in myself. The drink has called it into being, but does not know it, and can only repeat indefinitely, with a progressive diminution of strength, the same message which I cannot interpret, though I hope at least to be able to call it forth again…. I put down the cup and examine my own mind.
Desensitization of a sensory response is well known in the experimental literature and is often referred to as sensory adaptation. It is a very general phenomenon that occurs any time a given nerve cell or neural pathway is stimulated repeatedly. As discussed in chapter 8, the logic of this is that the nervous system is not constructed to register every sensory stimulus imposing on it, but only those that signal a sudden change from a former state. It is these that carry the most critical information; the same stimulus repeated carries less and less information, until a different stimulus occurs.
It seems clear that desensitization of the initial sensory mechanism takes place during Proust’s initial repeated attempts to conjure up the “truth.” However, a close reading of the text indicates that several other neural processes are likely also occurring. A second process is adaptation in the neural pathways that process the odor information to give rise to the odor perception; these would occur in the pathways of the brain flavor system. Adaptation may also occur in the pathways that link an odor perception per se to the systems underlying odor memory. Finally, there are the systems related to the narrator’s vision of the “truth”—the systems involved in the storage of the visual memories and their retrieval. The fading of the “truth” may thus be due to multiple mechanisms of adaptation, in addition to the desensitization of the receptors. But to assess this further, we need to ask what is the nature of the odor perception elicited by the potion.
Impulses in the fibers from the sensory neurons give rise to spatial patterns of activity within the first brain relay station, the olfactory bulb. These patterns are the smell images of the information carried in the smell molecules (chapters 4–10), which are projected to the olfactory cortex where they form a content addressable memory of the smell object (chapter 11), and are sent from there to the orbitofrontal cortex (chapter 12) to be combined with other sensory and motor systems to form the perception of smell and of flavor (chapters 13–18).
It is this flavor image that was recognized by Proust’s brain, at first only indistinctly, as being part of a more complex memory that initially seemed beyond recall. The flavor image of the tea-soaked madeleine is thus metonymic for the complex multisensory image of the town of Combray.
Smell, Emotion, and Memory Recall
The direct access of the smell pathway to these forebrain mechanisms is essential for understanding the nature of Proust’s olfactory-evoked experience. We have indicated the cortical mechanisms involved in Proust’s cognitive, perceptual response (chapter 18). This direct olfactory connection to the forebrain provides insight into the heightened degree of the emotional state evoked by the odor stimuli, the strength of the voluntary search for the missing “truth,” and the overwhelming quality of the “involuntary” memory finally brought forth.
The emotions evoked by the madeleine are central to the whole theoretical edifice of the madeleine episode:
An exquisite pleasure had invaded my senses, something isolated, detached, with no suggestion of its origin…. [T]his new sensation… had the effect which love has of filling me with a precious essence…. Whence could it have come to me, this all-powerful joy? I sensed that it was connected with the taste of the tea and the cake, but that it infinitely transcended those savours, could not, indeed, be of the same nature. Whence did it come? What did it mean? How could I seize and apprehend it?
Brain research can best provide insight into the question “Whence did it come?” We have explained how, from the olfactory cortex, the pathway for perception is directed toward the prefrontal neocortex. But the olfactory cortex also gives rise to multiple pathways that connect directly to the so-called limbic regions of the brain. These are the phylogenetically old regions of the brain that are involved in the mediation of both memories and emotions.
The key structures include the hippocampus, a central organizing node for single-event “episodic” memories, and the amygdala, which, in parallel with the orbitofrontal cortex, is involved in stimulus reinforcement association learning. The reader may refer to chapters 18 and 19 for an orientation to the places of these structures in the brain flavor system.
Proust believed that the recollection of Combray was an involuntary memory from the past, purer and nobler than memories recalled voluntarily. This notion became favored by a number of literary critics, including the writer Samuel Beckett. From these critics arose what has been called the salvationist idea that involuntary memories are kernels of pure truth, the hidden essences of ourselves, precisely because we cannot call them forth at will. But was Combray an involuntary memory? Proust’s text says otherwise:
I begin again to ask myself what it could have been, this unremembered state…. I decide to attempt to make it reappear. I retrace my thoughts to the moment at which I drank the first spoonful of tea. I rediscover the same state…. I ask my mind to make one further effort, to bring back once more the fleeting sensation…. I shut out every obstacle…. I compel it [my mind] for a change to enjoy the distraction which I have just denied it… for the second time I clear an empty space in front of it; I place in position before my mind’s eye the still recent taste of that first mouthful, and I feel something start within me… ; I do not know yet what it is, but I can feel it mounting slowly….
Undoubtedly what is thus palpitating in the depths of my being must be the image, the visual memory which, being linked to that taste, is trying to follow it into my conscious mind…. Ten times over I must essay the task….
And suddenly the memory revealed itself.
Thus, “the whole of Combray… sprang into being… from my cup of tea” is hardly accurate, yet many critics seem to have been misled by that dramatic pronouncement. More than a page of text, including the excerpts given here, goes into describing the enormous effort of voluntary recall and searching that took place, and shows that the actual sequence of events involved a considerable delay between the initial sensory input and the realization of the memory associated with it. Proust himself makes clear the delay that is involved, both at the time and much later, in Time Regained, when he muses once more on the power of the sensation and the time it took to identify the memory it conjured: “I had continued to savour the taste of the madeleine while I tried to draw into my consciousness whatever it was that it recalled to me.”
What are the neural mechanisms responsible for this enormous effort? From a neuroscientific point of view, the association between Proust’s intensely emotional state and his highly motivated state is understandable, because much experimental research has shown that the systems responsible for motivated behavior are part of the same deep brain systems involved in emotion. This is summarized in figure 18.2 and the discussion of the human brain flavor system in chapters 18 and 19. The smell and flavor inputs go directly to the forebrain motivational systems that mediated the will to search for the meaning behind the emotions.
Combray Recalled
Our analysis suggests that the recall of Combray is not as involuntary as Proust claims and as literary critics have assumed. From a neuropsychological point of view, the fact that it seems to come back at once and entire is also not surprising; our cognitive mechanisms have a gestalt quality in which we perceive and recall things as integrated wholes. The content-addressable memory mechanism of the olfactory cortex is designed specifically to recall a whole from a small part (chapter 11). Given a fragment, we (and Proust) have a strong tendency to make a whole. Combray comes back in its full visual embodiment in the normal way that most olfactory memories are realized.
Proust seems aware of this phenomenon, and in Time Regained he explores in minute detail how one can recognize a person’s face when shown only one part of it, such as the nose or mouth. The context for these observations of the “gestalt effect” is a costume party that Proust attends after a long absence from society. He wanders through the crowd, studying the disguised faces that are in effect doubly masked, by costume and by age. Beneath all the costumes are people he has known well yet does not immediately recognize because of his previous absence and because of their disguises, so he methodically studies the face of each person to find some familiar feature that will allow him to compose in his mind the whole face:
Thanks to a tiny fragment which still survived of the look that I remembered, I was just able to recognise the youth whom I had once met at Mme de Villeparisis’s tea-party…. I thus succeeded in identifying somebody, by trying to dismiss from my mind the effects of his disguise and building up, through an effort of memory, a whole familiar face round those features which had remained unaltered.
Proust proceeds around the room observing people and performing this same systematic approach to remembering them, a process that occupies some 40 pages of the novel.
Where in the brain does the memory trace of Combray reside? On the basis of work on other types of memory, we have a good idea that memories are stored in distributed systems in the brain. Given Combray’s iconic status, devising experiments to test for this kind of memory has been an irresistible temptation for a number of laboratories that use brain imaging techniques such as positron emission tomography (PET) and functional magnetic resonance imaging (fMRI).
An early study by Gerson Fink and his colleagues in 1996 using PET brain scans tested the idea that there is something special about memories from our own past, the kind of memory that Proust imagined. In this test, subjects sat in the scanner and listened to sentences relating emotional episodes in the life of someone they did not know, while the scanner hummed away recording their brain activity. The results were compared with the activity in subjects who read sentences relating to emotional episodes from their own past. The brain scans showed that the autobiographical episodes activated a specific set of regions. They were primarily in the right cerebral hemisphere, especially in the cortex of the temporal lobe including the hippocampus, parahippocampus, and amygdala as well as the more distant posterior cingulate, insula, and prefrontal areas on the right.
We have already met these regions as parts of the distributed systems for the perception of smell, and for the “images of desire.” All these parts of the brain are well known for their involvement in different kinds of memory tasks. Electrical stimulation of many of these areas in human patients by the neurosurgeon Wilder Penfield has long been known to be able to elicit long-term memories from childhood. The PET results suggest that autobiographical memory involves a subset of these regions mainly in the right hemisphere of the brain, the hemisphere commonly associated with more intuitive and nonverbal kinds of behavior.
On the Mechanism of Autobiographical Memory
Jay Gottfried and his colleagues at Northwestern University picked up this question with a study using functional brain scans in 2004. They were intrigued with the fact that an autobiographical memory does not recall just one sense, and that we can usually imagine the whole scene: the sounds of the ocean or the voices of the people there, or the smells in the wind or the food. When we recall a sight or a sound, regions involved in higher-order processing of vision and hearing are reactivated, as if the subjects are internally viewing the memory traces of the original sights and sounds. This is very similar to “mental imagery.”
Gottfried and his colleagues asked whether they could demonstrate a recollection in the brain if it involved linking two different senses. They tested subjects by exposing them to a smell (either pleasant or unpleasant) followed by a picture of an object (table, lamp), during which time they were asked to imagine a story or some kind of link between the two. Later, the subjects were placed within the brain scanner, exposed to a variety of objects, and asked which ones reminded them of a smell. Behaviorally, the subjects were best at identifying the association of a visual object with a pleasant smell. The brain activity patterns were dramatic in showing prominent activity in the olfactory cortex, even though the subjects were merely remembering the smell, not actually sniffing it.
How does the brain bind together reactivation of a visual object in the visual cortex with representation of a smell in the olfactory cortex? It was known that in the case of vision and hearing, the memory traces are stored in relation to the different senses but bound together in the hippocampus. Gottfried and colleagues therefore hypothesized that this could also occur in the case of vision and smell. Since the olfactory cortex has no direct connection with the visual cortex, the connection must occur somewhere else; the prime suspect is accordingly the hippocampus and neighboring parahippocampus in the temporal lobe, the same regions found in the study by Fink and his colleagues. The authors conclude:
[T]he hippocampus binds a distributed trace maintained across sensory-specific regions. Such a system preserves the integrity of the original engram and enables the access by partial or incomplete cues, lending flexibility and adaptability to the memory system…. [I]t is less likely that sensory elements of the original trace would be distributed in higher-order olfactory areas, such as [the orbitofrontal cortex] or cingulate cortex, where sensory fidelity is inevitably compromised through progressive synaptic convergence and divergence. This factor may help to explain the corresponding absence of retrieval-related activity in these particular regions.
This is getting close to telling us what Combray “looked like” in Proust’s brain. In these experiments a visual object was used to recall a smell; for Proust, a retronasal smell recalled a visual scene. But the principles appear to be similar: an internal reactivation of the distributed sensory regions bound together by their connections to the hippocampus. Rather than a perfect memory suddenly bursting forth, it appears that Proust was describing positively reinforcing sensory stimuli from a childhood experience, stored in their respective central sensory representations, bound together by their connections to the hippocampus, and reaccessed, beginning with partial flavor cues, as a unified internal image or object by the brain mechanisms of attention, motivation, and emotion.