Please join us for a series of research talks from students in the BCS graduate program. 

Date/Time: Friday, March 8th from 4pm - 6pm

Location: On Zoom https://mit.zoom.us/j/96094784666

Aida Piccato, Year 4 Graduate Student in Mehrdad Jazayeri’s Lab

Title: Neural correlates of visual long- and short-term memory in the primate prefrontal cortex

Abstract: Primates exhibit a rich repertoire of behaviors dependent on visual memory, or memory of objects and their locations. Visual memory spans multiple timescales - visual short-term memory (vSTM) stores information on the order of seconds, and is critical for perceptual decision-making. Visual long-term memory (vLTM) stores information over much longer timescales, and is the basis of our knowledge of the visual world. Functional and anatomical distinctions between these two types of memory have inspired cognitive models of memory in which vLTM and vSTM are considered dependent on distinct retrieval and storage mechanisms. We have developed a novel behavioral paradigm suitable for rhesus macaques to examine the relationship between in short- and long-term memory representations in the dorsolateral prefrontal cortex. We have trained one monkey on the task, and begun recording from the ventrolateral prefrontal cortex. We find that dLPFC neurons exhibit varied selectivity to memory type and content. We then test cognitive theories regarding the relationship between neural states in vLTM and vSTM. We find that  vSTM and vLTM memory content are encoded by parallel, offset sub spaces in this region.  We propose possible memory architectures consistent with these findings. 

Gal Raz, Year 5 Graduate Student in Rebecca Saxe’s Lab

Title:  A rational model of infant looking time

Abstract: From birth, humans learn actively. Developmental psychologists have long capitalized on this fact, probing infants' mental representations through their looking behavior. Despite being a key measure, we do not have a rigorous, formal framework for why infants look longer at some stimuli than others. To address this, we developed a rational learning model that decides how long to look at sequences of stimuli based on its expectation to gain information. The model captures key patterns of looking time documented in the literature. By using a CNN-derived embedding space, the model can operate on raw images and generate novel predictions for previously untested stimuli. We validate these predictions by collecting two large infant looking time datasets (N = 145), and comparing model and infant behaviors. We argue that our model is a general and interpretable framework for the rational analysis of looking time.

Margaret Schroeder, Year 5 Graduate Student in Guoping Feng’s Lab

Title: A transcriptomic atlas of astrocyte regional heterogeneity across developmental stages in mouse and marmoset brains

Abstract: Astrocytes are an abundant class of glial cells with critical roles in neural circuit assembly and function. Recent single-cell and earlier bulk RNA sequencing studies have demonstrated significant transcriptomic heterogeneity among astrocytes, particularly across brain regions. However, the developmental trajectory of this heterogeneity and its conservation across species requires further systematic study. To this end, we used single-nucleus RNA sequencing to characterize the molecular diversity of brain cells across developmental stages (late embryonic, neonate, early and late adolescent, young adult, and late adult) and four brain regions (prefrontal cortex, motor cortex, thalamus, and striatum) in the mouse and marmoset brain. We made an effort to employ consistent dissection strategies, sample preparation protocols, and sequencing technology across groups to enable clean comparison of biological differences with reduced technical artifacts, which are extremely difficult to remove in silico. Our analysis of over 150,000 single astrocyte nuclei revealed striking regional heterogeneity among astrocytes, particularly between telencephalic and diencephalic regions, at all developmental time points surveyed in both species. Top differentially expressed genes between telencephalic and diencephalic astrocytes, many of which were conserved across species, implicated calcium signaling and glutamatergic synaptic transmission, suggesting that regional astrocytes may be molecularly specialized to support their local neuronal circuits. Though astrocytes were already regionally patterned in late embryonic stages, portions of this region-specific astrocyte gene expression signature changed over postnatal development: we found several genes which were regionally differentially expressed in newborns but not adolescents or adults, and vice-versa. More broadly, this cross-species, cross-development, cross-region molecular profile of brain cells using consistent experimental and computational methodology is a valuable resource for the field.

Sadie Zacharek, Year 4 Graduate Student in John Gabrieli’s Lab

Talk Title: Prediction in neurodevelopmental and neuropsychiatric disorders

Abstract: Characterization of neurodevelopmental and neuropsychiatric disorders has greater translational value if we can make valid predictions about disorder pathology. For instance, predicting which at-risk individuals will go on to develop a disorder or predicting which treatment is most likely to benefit an individual patient. 3 ongoing projects centered on prediction will be discussed. The first, a longitudinal study of preclinical preschool-aged children at familial risk for ADHD which aims to disentangle the causes vs. consequences of ADHD. The second, a clinical trial for adult social anxiety disorder which aims to identify neurobiomarkers predictive of treatment success. Finally, a treatment study of familial factors involved in the genesis and maintenance of childhood depression, parent-child brain similarity, and changes with family-focused treatment. These studies each aim to uncover predictive information about each disorder that goes beyond characterization to clinical utility.

Yudi Xie, Year 5 Graduate Student in James DiCarlo’s Lab

Talk Title: Sensory representation mismatch explains working memory capacity limitation

Abstract:  The limited capacity of the brain to retain information in working memory has been well-known and studied for decades, yet the root of this limitation remains unclear. Here we built sensory-cognitive neural network models of working memory that perform tasks using raw visual stimuli. Contrary to intuitions that working memory capacity limitation stems from memory or cognitive constraints, we found that pre-training the sensory region of our models with natural images imposes sufficient constraints on models to exhibit a wide range of human-like behaviors in visual working memory tasks designed to probe capacity. Examining the neural mechanisms in our model reveals that capacity limitation mainly arises in a bottom-up manner. Our models offer a principled and functionally grounded explanation for the working memory capacity limitation in terms of sensory representation mismatch. This work highlights the importance of developing models with realistic sensory processing even when investigating memory and other high-level cognitive phenomena.