The Neural Foundations for Memory and Social Cognition in the Human Brain

National Institute Of Mental Health

Investigators

Linked publications, trials & patents

Abstract

During this past year we have concentrated our efforts on addressing outstanding issues concerning the three major divisions of memory: semantic or conceptual memory composed of our knowledge about things and ideas, explicit or autobiographical memory composed of recollections of our life events, and priming, a form of implicit memory that underlies our ability to recognize objects fast and efficiently. One of the most prominent models of the functional neuroanatomy of semantic memory posits that the more anterior region of the temporal lobes functions as a domain-general convergence zone, or hub, for all semantic information. This view is at odds with evidence from our lab and others that indicate that the anterior temporal lobe (commonly referred to as the ATL) comprises multiple anatomic and functional subdivisions. There have been two major obstacles to using functional magnetic resonance imaging (fMRI) to identify the precise functional organization of the ATL: the difficult choice of stimuli and tasks to activate, and dissociate, specific regions within the ATL, and poor signal quality because of magnetic field distortions in this region of the brain. To circumvent these difficulties, we developed a data-driven parcellation routine and an imaging sequence optimized to enhance signal in the ATL. Our analysis revealed that the ATL is composed of at least 34 distinct functional regions. In addition, the region most often claimed as the location of the semantic hub was found to be part of a domain-specific network associated with face and social processing, rather than a domain-general semantic hub. These findings offer a fine-grained functional map of the ATL and offer an initial step toward using more precise language to describe the locations of functional responses in this heterogeneous region of human cortex (Persichetti et al., Journal of Neuroscience, 2021). Several of our other studies of semantic memory have focused on the relationship between the neural systems for representing higher-order conceptual information and lower-level sensory information. We recently established, using high field (7-tesla), high-resolution fMRI, that taste information (sweet, salty, and sour tastes) is represented in gustatory cortex by a population code, rather than by a taste-specific spatial map (Avery et al., Journal of Neuroscience, 2020). Building on this finding we asked whether this brain region also contained a population code for representing the inferred tastes associated with different types of foods. Specifically, we asked whether simply viewing pictures of foods defined primarily by their dominant taste (e.g., pictures of candy, pretzels, lemons) would activate the same region of gustatory cortex active when experiencing these tastes, and whether these inferred tastes were represented by a population code rather than by a spatial map. Again, using ultra-high resolution fMRI at high magnetic field strength, we found that this was indeed the case. Viewing food pictures triggered the automatic retrieval of specific taste quality information associated with the depicted foods and that this information was represented by a population code (Avery et al., Proceedings of the National Academy, 2021). These results also showed how higher-order inferences derived from stimuli in one modality (i.e., vision) can be represented in a region of the brain typically thought to represent only low-level information about a different modality (i.e., taste). We are also interested in how memory for our life events and occurrences (autobiographical memory) are represented in the brain. For several decades, the role played during autobiographical retrieval by the hippocampus one of the central brain regions in memory functioning has been hotly debated. One model (The Standard Consolidation Model) asserts that via a time-dependent process of cortical consolidation, the hippocampus is no longer required for successful retrieval of remote events (e.g., during childhood). In contrast, others have argued that the critical determinate of hippocampal involvement is not how long ago the event occurred, but the precision and detail of the memory. In this view, the hippocampus is always required for vividly recalling memories (The Multiple Trace hypothesis). Although the models make clear predictions, distinguishing between them has been complicated by the tendency for memories to become more schematic and less detailed over time, such that age and vividness are confounded. The debate for and against temporally graded activity in the hippocampus rest largely on the subjective phenomenology of retrieved memories. In this regard, fMRI studies of autobiographical memory have suffered in particular because of their reliance upon covert, or silent, recall because of concerns about in-scanner head motion. We overcame this problem by capitalizing on recent advances in fMRI acquisition and analysis that allowed our participants to overtly recall their recent and remote memories during scanning. In accordance with the predictions of the Standard Consolidation Model, we found that the hippocampus became less and less active the more remote the retrieved memory, with no hippocampal activity detected for memories recalled from the most distant past. Moreover, by careful analysis of the actual verbal memory reports generated by our participants we were able to show that this result was not due to the amount or quality of the details recalled (Gilmore et al., Proceedings of the National Academy, 2021). The same verbal reports provided a means of examining how autobiographical memories are retrieved in real-time, as changes in neural activity could be correlated with changes in content as it was being described. The results indicated rich, complex retrieval dynamics, whereby domain-specific cortices are transiently reactivated to support moment-to-moment re-experiencing of an event (Gilmore et al., Journal of Neuroscience, 2021). Finally, our studies of implicit memory have focused on the behavioral and neural underpinnings of a powerful form of learning known as priming. It has longed been recognized that our ability to identify a stimulus improves with repetition (repetition priming), whereas neural activity decreases (repetition suppression). Multiple models have been proposed to explain this unexpected brain-behavior relationship (i.e., improved performance coupled with decreased brain activity). Some prominent models predict that this phenomenon is underpinned by changes in neural connectivity between specific brain regions (Synchrony and Predictive Coding models), others predict changes in the latency of neural responses (Facilitation model), and others in a change in the relative similarity of neural representations (Sharpening model). In a recent study we tested these predictions with fMRI while participants named repeated and non-repeated objects. While we found partial support for predictions of the Facilitation and Sharpening models, the data were most consistent with the Synchrony model, with increased coupling between specific regions of the brain for repeated objects that correlated with the magnitude of the behavioral facilitation associated with object repetition (priming) (Gotts et al., Communications Biology, 2021). Surprisingly, increased regional coupling and the neural phenomenon of repetition suppression varied independently, a finding we will pursue in future studies.