Semantic Memory for Food and Brain Correlates

Semantic memory stores knowledge about different types of objects: plants, animals, vehicles, utensils, conspecifics and food, among the others. Our ability to quickly recognize and categorize an object when we encounter it depends upon having experienced that object before and on semantic knowledge integrity. Semantic memory is one the most resilient cognitive abilities, it is less prone to interference than episodic memory and more declines slowly. The interest in how semantic memory is organized traces way back, however a great impulse was provided by the first systematic neuropsychological observations of patients with category specific recognition deficits. However, this debate is far from being resolved. In my dissertation, I will show how the study of food as a semantic category is extremely suitable to shed light on the organization of semantic knowledge. The thesis is organized as follow. In Chapter 1, I will first define semantic memory, focusing on its characteristics, such as its relationship with experience, its resilience to cognitive decline and its neural correlates, and on how it has been studied by neuropsychologists. In addition, I will review the studies on the food category, focusing on some intrinsic dimensions such as the level of transformation. Chapter 2 includes Study 1, in which I have investigated the organization of semantic memory by using food (natural and transformed) and non-food (living and on-living things) in a group of patients suffering from temporal lobe atrophy (Alzheimer’s disease, PPA and FTD) and healthy controls, using Voxel Based Morphometry and DTI. Results have shown that food breaks down in natural and transformed, and that this parsing mirrors that of living and non-living things, thus strongly supporting the Sensory-functional model of semantic knowledge. Chapter 3 contains Study 2, in which I have explored the relationship between semantic memory and experience. I collected information about life-long eating habits as a proxy of long-term experience with specific foods as well as information about semantic memory of food in participants of different ages (36 – 108 years old). Results support the hypothesis that semantic memory is modulated by experience. In Chapter 4, the focus of Study 3 is on episodic memory. Here I investigated whether the difference between semantic memory for natural and transformed food highlighted in Study 2 extends also to episodic memory, and whether the animacy effect - a facilitation to remember living exemplars - holds for food as well. Specifically, I administered a recognition memory task to the same participants of Study 2, to a group of young participants and to patients with Alzheimer’s disease, PPA and FTD. I found that young adults had better recognition memory for transformed foods compared to natural foods. This difference disappeared in centenarians, consistently with Study 2, and in patients. The natural/transformed distinction appears susceptible to decay only in the presence of a high degree of episodic memory impairment. Finally, with Study 4, described in Chapter 5, my aim was more translational, that is, to test whether a deficit in semantic memory for food could lead to specific eating disorders. This study empirically establishes the behavioural and neural correlates of abnormal changes in eating habits in dementia and their relationship with semantic memory. In this thesis, I have shown that natural and transformed food do have different neural correlates, and that they are differently represented in semantic memory. By drawing together evidence from my studies and from studies of others I was allowed to propose a comprehensive model of semantic knowledge. Additionally, in my thesis I showed how food can be employed to study the organization of semantic knowledge, the way in which semantic knowledge is shaped by learning and experience, and its effect on behaviour.