Search

Tilføj Ida Maria Mathilde Thomsen ...

Visual short-term memory has a surprising, limited capacity (e.g., Sperling, 1960) and is often described as a being fixed to a magic number (Miller, 1956; Cowen, 2001). Nevertheless, Alvarez and Cavanagh (2004) have suggested that this limitation vary based on stimulus complexity. Also, Jackson and Raymond (2005) made a distinction between physical and perceived complexity. Here we present three studies suggesting that short-term memory capacity is depended on long-term memory representations (Brogaard & Sørensen, in press). In a study by Sørensen and Kyllingsbæk (2012) we demonstrate that memory capacity systematically increases with degree of training and familiarity compared with a baseline condition consisting of linedrawings (Snodgrass & Vanderwart, 1980). These results are later replicated and extended to a different paradigm (Dall, Watanabe, & Sørensen, 2016) demonstrating that training languages modulate the efficiency of representation in short-term memory. Similar results have later been reported for non-language categories link Pokémon cartoons by Xie and Zhang (2017). In the final study presented here we explore the dynamics between physical and perceived complexity in expert observers and demonstrate that both speed of processing and memory capacity is driven by perceived complexity independent of physical complexity (Dall, et al., 2021). Our results align with a view where initial sensory evidence is matched to long-term memory representations as they are selected for encoding in short-term memory (Bundesen, 1990), and as long-term memory templates become stronger with familiarity processing becomes faster and more efficient (Brogaard & Sørensen, in press).

How and what information we store in our Visual Short-Term Memory (VSTM) has been extensively explored, particularly in recent years (e.g., Bundesen, 1987; Schurgin et al., 2020; Zhang & Luck, 2008). Extending Sørensen and Kyllingsbæk (2012) we present a series of behavioural experiments that demonstrate how expertise influences what we see (Experiment 1-3) and in what order it is perceived (Experiment 4). In the first experiment Dall, Watanabe, & Sørensen (2016) examined how expertise influenced VSTM for Japanese Hiragana in participants at different levels of language expertise. Secondly, Dall et al. (2021) studied the relationship between the physical and perceived complexity (Jackson & Raymond, 2005) using Chinese characters and thirdly how stimuli specific expertise influenced both amount encoded and precision of stimuli in VSTM (Dall & Sørensen, 2019). Finally, Experiment 4 explore how expertise affect asynchronously presented stimuli in VSTM (Dall & Sørensen, in press).We find that expertise influences VSTM capacity to the point where it is not influenced by visual complexity, nevertheless it does not seem to influence the precision of representation in VSTM. In addition, we find a bias towards remembering high expertise stimuli when presented alongside low expertise stimuli.

Several theories have been proposed on how Visual Short-Term Memory (VSTM) encode stimuli (e.g., Bundesen, 1987; Schurgin et al., 2020). Theory of Visual Attention (Bundesen, 1990) assumes that attention functions as a race into a limited VSTM. Here we examine how two different classes of stimuli are processed if one is presented with a slight head start. This relationship was examined in a whole report experiment where we presented eight stimuli in a memory array. Three conditions were implemented consisting of either letters, pictures, or a mixture of letters and pictures. In two additional critical conditions we presented mixed stimuli, providing a 50 ms onset advantage to either pictures or letters. The onset was controlled with pre- and post-masking of the memory array. Participants were encouraged in all conditions to report as many stimuli as possible regardless of type. Previously research have demonstrated a VSTM advantage in capacity for letters (Sørensen & Kyllingsbæk, 2012; Dall et al., 2016). Therefore, participants are expected to have letter advantage in processing, even when pictures are provided a head start in onset. By varying exposure durations (5-200 ms) we explore the processing race into VSTM and find on synchronises trials that participants remember more pictures than letters at lower exposure durations. However, at higher exposure durations participants demonstrate a letter advantage. This seems to suggest that pictures have an early processing advantage which is progressively reduced on the later exposure rates.

Humans have a limited visual short-term memory capacity (K), which is further reduced when processing certain stimuli (e.g., unfamiliar, visually complex, or low discriminability). To examine how these factors influence the bottleneck in attentional and short-term memory, we conducted a series of studies. Experiment 1 (Dall, Watanabe, Sørensen, 2016) examined how expertise modulate K in Japanese hiragana. Experiment 2 (Dall, Wang, Cai, Chan, Sørensen, 2021) divided Chinese characters into four groups based on frequency of use (high vs. low) and stroke count (high vs. low) and how attention is influenced by number of features and stimuli specific expertise. Experiment 3 (Dall, Chan, Sørensen, in review) used colours to examine how stimuli specific expertise influence contentious categories. We extracted attentional components by varying the exposure duration and modelling the data with the Theory of Visual Attention in all three experiments, measuring K and processing speed (C).All three papers showed that stimuli specific expertise improved the K. This increase was unrelated to both stimulus complexity (Experiment 2) and precision of the memory (Experiment 3). Nor did visually complexity influence C of even complex and infrequently Chinese characters (e.g., 鹿), whereas C was modulated by familiarity (Experiment 2).This seems to suggest that the bottleneck in VSTM is influenced by stimuli specific expertise and that this increase cannot be explained as increasing the precision of the stored items.

Visual complexity of an object can vary in different degrees, and if we have a limited short-term memory capacity (e.g. Cowan, 2001), how does memory deal with objects of high visual complexity? While early studies demonstrated that visual complexity increase the load on both processing time and memory capacity (Alvarez & Cavanagh 2004), later studies indicate that memory capacity is highly influenced by stimulus specific expertise (Sørensen, & Kyllingsbæk, 2012). Nevertheless, complexity can be divided into two types based on either visual elements or familiarity. The current aim is to examine the relationship between the visual complexity and what we term the perceived complexity. In an experiment we used Chinese characters (radicals) divided into four groups based on visual complexity or stroke count (high vs low) and perceived complexity or word frequency (high vs low). We examined this using a whole report paradigm (Sperling, 1960) with varying exposure rates and modeled the results according to the Theory of Visual Attention (Bundesen, 1990). This procedure allows us to isolate specific attentional components and the results reveal that for experienced observers differences in processing speed and memory capacity is driven solely by the perceived complexity of an object.

Recently research has shifted from examining discrete categories (e.g. letters) to continuous categories (e.g. colours). While studies have shown that stimuli specific expertise influence discrete categories, there are little research into how it influences continuous categories.The current study comprises of a whole report experiment with two conditions: discrete and continuous. We divided participants into two groups based on the degree of colour expertise, normal colour expertise and colour experts. By varying stimulus exposure durations, we can fit data using the TVA in order to isolate specific components of attention like visual short-term memory capacity (K), speed of processing (C), and the threshold of perception (t0). Results reveal no significant group differences in the discrete letter condition, but that colour expertise improve K in the continuous category, leaving C and t0 unaffected, indicating that continuous categories improve from expertise similar to discrete.

Several studies discuss the influence of complexity on the visual short term memory; some have demonstrated that short-term memory is surprisingly stable regardless of content (e.g. Luck & Vogel, 1997) where others have shown that memory can be influenced by the complexity of stimulus (e.g. Alvarez & Cavanagh, 2004). But the term complexity is often not clearly defined. Sørensen (2008; see also Dall, Katsumi, & Sørensen, 2016) suggested that complexity can be related to two different types; objective and subjective complexity. This distinction is supported by a number of studies on the influence of expertise (e.g. Dall, et al., 2016). We will present a paradigm testing the proposed distinction using specific isolation of attentional components (see Bundesen, 1990; Sørensen, Vangkilde, & Bundesen, 2015). We propose that objective complexity can be manipulated through the number of strokes in Chinese characters. On the contrary expertise or word frequency may reflect what could be termed subjective complexity, as this relate directly to the individual mental categories established. This study will be able to uncover more details on how we should define complexity of objects to be encoded into short-term.