Advancing Cognition through Adversarial Collaborations Seeking Synthesis: A Paradigm Shift?
28-30 May 2020 | Royal Society of Edinburgh, Edinburgh, UK
As in many scientific endeavors, and indeed in broader society, the research literature on cognitive psychology is characterized by debate, and debate is fuelled by contrasting theoretical assumptions and contrasting data patterns. This might be viewed as a ‘between-groups’ approach to our fields of research. That is, the opponents in the debate tend to work separately in different laboratories, and groups of researchers tend to comprise like-minded individuals. A finding is established with evidence from one group of researchers, only to be challenged by another research group who might use a different paradigm, and who offer evidence that the original finding by the other group was flawed. The clear goal for each group is to win the debate. This approach to science can lead to advances in knowledge in our field, and in some areas of cognition very substantial progress has been made. But it can also lead to perpetuating seemingly endless debates without making any real progress.
A famous article by Newell (1973) ‘You can’t play 20 questions with nature and win’ listed 24 binary oppositions that were the subject of several years, if not decades of ongoing debate in the cognitive psychology literature at that time. Despite the generation of large volumes of data, many of those binary oppositions remain unresolved after nearly 50 additional years of research. Some have simply become unfashionable, but many remain the focus for contemporary debates at the leading edge of research that might be considered as ‘stalled’ by debate, for example, trace decay versus interference in forgetting, single memory versus dual memory (STM-LTM), serial versus parallel processing, conscious versus unconscious processing, to name but a few. These, and a wide range of additional contemporary debates remain important issues for cognition, with no obvious path to theoretical or empirical resolution that might be universally accepted as a genuine advance in scientific understanding. One possible reason for this apparent lack of progress might be that opposing views become entrenched with particular groups, and new theories are developed, or old theories are reinvented because this is a route to academic career development rather than to genuine knowledge advancement. A second possible reason is that much research on cognition has continued with the approach that Newell saw as a barrier to progress, to argue for one or other side of a chosen binary, or multi-way divide.
What Changes in Cognition during Working Memory Training?
Presenter(s): Susanne Jaeggi, University of California, Irvine, Thomas Redick, Purdue University, Monika Melby- Lervåg, University of Oslo, Susan Gathercole, University of Cambridge
This topic has a focus on whether, and if so why, training on a cognitive task, or a set of tasks results in enhanced performance of untrained tasks, and whether any such effect of training applies only to tasks that are similar to the training task(s) or offers more general enhancement of mental ability. The general issue of cognitive training was studied extensively during the second half of the 20th century, with early work in the 1940s (e.g. Miller, 1947), and in later work on methods for training (e.g. Gopher et al., 1989; Moray et al., 1986; Schneider, 1985), on transfer of training (e.g. Povenmire, & Roscoe, 1973; Singley & Anderson, 1989), and the development of cognitive expertise (e.g. Ericsson et al., 1980; Ericsson et al., 2018). The contemporary, leading edge debate on working memory training tends not to refer to that previous research, and has been stimulated by studies suggesting that training working memory leads to an enhancement of general intelligence and can counter the impact of cognitive decline with age and in other disorders such as ADHD (e.g. Jaeggi et al. 2008; 2018; Klingberg, 2002). Other studies have suggested that these findings may be unreliable (Melby-Lervåg & Hulme, 2015; Marcelle et al., 2018; Redick, 2019; Shipstead et al., 2012), with some recent studies suggesting that new cognitive skills show improvements with training, but well established cognitive skills do not (Gathercole et al., 2019). The debate raises important scientific questions about whether or not general mental ability can be enhanced, but also has important and widespread implications beyond the laboratory for implementation of educational policy and interventions, together with commercial pressures.
What Are the Causes of Long-Term Forgetting?
Presenter(s): Michael Anderson, University of Cambridge, Aya Ben-Yakov, MRC Cognition and Brain Sciences, Lili Sahakyan, University of Illinois at Urbana-Champaign, Jeffrey Zacks, Washington University in St. Louis
This topic has a very long history, going back to the work of Ebbinghaus in the 19th century. However, it is at the core of important contemporary debates. There are multiple issues unresolved, notably regarding the influence of extrinsic factors in contrast with intrinsic factors on forgetting of real-life events, and the role of the hipocampus (see topic 3). Important, contemporary research on extrinsic factors has focused on event boundaries, for example 'Walking through doors causes forgetting' (O’Rear & Radvansky, 2019; Radvansky & Copeland, 2006; Radvansky et al., 2015). This shows forgetting of information carried across event boundaries, but better memory for information at event boundaries or that can be separated by event boundaries (e.g. Kurby & Zacks, 2008; 2018; Pettijohn et al., 2016). It is also unclear as to whether the extent to which there is hippocampal activity correlates with on or off signals for event boundaries (e.g. Ben-Yakov, 2013; Ben-Yakov & Henson, 2018). Major intrinsic factors under debate include whether long-term forgetting arises from material-based interference at encoding or retrieval, disruption of attention, or a failure of consolidation (e.g. Craig & Dewar, 2018; Dewar, Cowan, & Della Sala, 2007; Dewar et al., 2009; 2010; Ecker et al., 2015). Other key intrinsic factors include retrieval-induced and directed forgetting (e.g. Buchli, Storm, & Bjork, 2016; for a contrasting view see Akan & Sahakyan, 2018), and retrieval suppression (e.g. Taubenfield, Anderson, & Levey, 2019). How these various extrinsic and intrinsic factors might interact and their relative contributions to forgetting tends not to be addressed yet the combination of these factors has important theoretical implications, as well as major implications for understanding real-life healthy and pathological forgetting, for example in post-traumatic stress disorder (e.g. Hulbert & Anderson, 2018). It is anticipated that discussion within this topic will inform discussion within topic 3 below, in addition to addressing the key debates within the cognition of forgetting.
What Does the Hippocampus Do?
Presenter(s): Richard Allen, University of Leeds, Masud Husain, University of Oxford, Eleanor Maguire, University College London, Sergio Della Sala, University of Edinburgh
The hippocampus has long been associated with the process of encoding and retrieving episodic traces (Scoville & Milner, 1957; Squire, 1992), but in more recent studies has been shown to be important for a wide range of cognitive functions including spatial navigation, mental imagery, scene perception and scene representation (e.g. Maguire & Mullally, 2013; Maguire, Nannery, & Spiers, 2006), imagining the future, and mind wandering (for an overview see Ekstrom & Ranganath, 2018). The hippocampus has also been implicated in short-term memory (e.g. Hartley….Husain, et al., 2007; Zokaei ... & Husain, 2019). In contrast, patients with damage to the hippocampus have been shown to have normal short-term/working memory despite dense anterograde amnesia (e.g. Allen, Vargha-Khadem, & Baddeley, 2014; Baddeley, Allen, & Vargha-Khadem, 2010; Kapur & Logie, 2003; Parra…..& Della Sala, 2015). Moreover, in healthy adults, the hippocampus has been shown not to be involved in temporary binding of colours and shapes for conjunctive binding (e.g. Parra, Della Sala, Logie, & Morcom, 2014), although it is clearly involved in forming associations in relational binding (e.g. Hannula & Ranganath, 2008). Recently, Dalton, Zeidman, McCormick, and Maguire (2018) have argued that it is too simplistic to consider the hippocampus as an homogeneous structure, and proposed that there are different circuits within the hippocampus, each having a different function. We anticipate that this discussion will be informed by the discussion on topic 2 (long-term forgetting).
Is the Capacity of Visual Working Memory Constrained by Number of Slots, by Precision, or by Activation of LTM Memory Traces?
Presenter(s): Nelson Cowan, University of Missouri, Columbia, Masud Husain, University of Oxford, Robert Logie, University of Edinburgh, Edward Vogel, University of Chicago
In an extension of the Baddeley & Hitch (1974) multiple component framework of working memory, Baddeley (1986) introduced the concept of a visuo-spatial sketchpad that was thought to be a system for supporting visual imagery and the temporary retention of visual and spatial properties of visually presented material. However, Phillips & Christie (1977) provided evidence that memory for abstract random square matrix patterns is limited to a single item, although that item may be a pattern that varies in complexity, and the number of elements within a single pattern also constrains capacity (Wilson, Scott, & Power, 1987). This, and later work (e.g. Logie, 1995; Borst, Niven & Logie, 2012) suggested that visual short-term memory might be a separate small capacity system for a single visual array. This was in contrast with visual imagery that has a large capacity for meaningful visual representations of objects and scenes (e.g. Paivio, 1971; See also topic 3 above). The debate that dominates contemporary research is concerned primarily with how the capacity for temporary retention of a single array is constrained. This was stimulated by Luck and Vogel (1997) who provided evidence that the capacity limit was around 4 items, and that these items could be either single features, such as a line orientation or a color, or could be integrated objects such as a green line at 45 degrees. Their conclusion was challenged by Wheeler and Treisman (2002) who argued that the capacity was limited by number of individual features, not by integrated objects. Cowan, Blume, and Saults (2013) argued that capacity of working memory is limited to around 4 items regardless of the type of material, and whether one or both attributes of a given object are retained, depends on how attention is allocated during encoding of the stimulus. Shimi & Logie (2019) argued for a specific visual store that can retain a single visual array (Phillips and Christie, 1977), limited to four items in the array. Rhodes, Cowan, Hardman, and Logie (2018), and Adam, Vogel, and Awh (2017) also have offered evidence that performance in tasks that test working memory for a visual array is influenced by informed guessing about the visual features of items. In contrast, Bays and colleagues (Bays, Catalao, & Husain 2009; Bays, Wu, & Husain, 2011; Ma, Husain, & Bays, 2014; Taylor & Bays, 2018) argued that the limitation is based on how much visual detail can be retained, not on number of items, and that the more detail in a stimulus, the less precise is the visual representation.