Working Memory for Spatial Sequences: Developmental and Evolutionary Factors in Encoding Ordinal and Relational Structures

Abstract

Sequence learning is a ubiquitous facet of human and animal cognition. Here, using a common sequence reproduction task, we investigated whether and how the ordinal and relational structures linking consecutive elements are acquired by human adults, children, and macaque monkeys. While children and monkeys exhibited significantly lower precision than adults for spatial location and temporal order information, only monkeys appeared to exceedingly focus on the first item. Most importantly, only humans, regardless of age, spontaneously extracted the spatial relations between consecutive items and used a chunking strategy to compress sequences in working memory. Monkeys did not detect such relational structures, even after extensive training. Monkey behavior was captured by a conjunctive coding model, whereas a chunk-based conjunctive model explained more variance in humans. These age-and species-related differences are indicative of developmental and evolutionary mechanisms of sequence encoding and may provide novel insights into the uniquely human cognitive capacities.

Efecto de la contigüidad espacial sobre el aprendizaje de secuencias de posiciones

“…Relationships can be established between events, between an event and its place of occurrence, or between these and adjacent or context cues, as suggested by the associative hypotheses presented in the introduction (Hurlstone et al, 2014;Kao et al, 2020, Lindsey, 2019Lindsey & Logan, 2021). As Zhang et al (2022) have pointed out, it is possible that this type of relationship in the human case is possible from language, which would explain the differences in performance in this type of task between humans and non-humans. In this study, there was no record or data that could constitute evidence of the participation of language in an eventual process of segmentation of the sequences, so direct evidence would be required in this regard in subsequent investigations.…”

“…The repetition of the sequence until its correct reproduction could have favored this progressive learning. The use of fragmentation or segmentation strategies have been proposed by other authors, mainly in humans but also, although to a lesser extent, in non-humans (macaques) (Botvinick et al, 2009;Miller, 1956;Zhang, et al, 2022). Lindsay and Logan (2021) have recently stated that an element of the sequence can be associated with several subsequent components because they are recognized as belonging to the same group of events.…”

“…Scarf et al (2011) found the same effect using 5-item sequences. Zhang et al (2022) made a comparative study between humans (adults and children) and monkeys (macaques) using an immediate serial recall task of 3, 4, 5 or 6 positions. As the length of the sequence increased, recall accuracy decreased in all cases, being more accentuated in the case of the monkeys in the sequences from 3 positions forward.…”

Shared structure facilitates working memory of multiple sequences

“…Factorization representation is thought to help fast generalization of a previously learned structure to new contents (Sheahan et al, 2021; Zhou et al, 2020). Indeed, the ability to spontaneously perceive relational structures is posited to signify the major distinction between human and nonhuman primates (Dehaene et al, 2015; H. Zhang et al, 2022). Meanwhile, previous modelling works also suggest that higher-order structures incorporated in WM would serve as constraints on individual-item representations to reduce representational uncertainty (Brady & Tenenbaum, 2013; Ding et al, 2017).…”

Feature based analysis of endometriosis using machine learning

“…They are: i) one-hot encoding and ii) ordinal encoding. The ordinal encoding technique [11] was adopted for the given dataset as it contains categorical values. The ordinal encoding was applied to the independent attributes of both the training and testing datasets.…”