Understanding Cognition: A Behavior Analyst’s Research Story

Guest Blog by Dr. Ian Tyndall: 

Functional Behavioural Science Laboratory, Department of Psychology, University of Chichester, UK


Thank you to Ruth Anne Rehfeldt for the kind invitation to write this blog. My research journey in behaviour analysis has not been a typical one and in a sense I have always felt, and still do, somewhat on the margins. During my undergraduate degree in Psychology at NUI, Galway in Ireland, behaviourism and the terms classical and operant conditioning were seen as something to be covered in the first year of the curriculum, in a similar vein to psychoanalysis, as a box to be ticked but never to be mentioned again for the remainder of the degree as it was ‘old hat’ and not part of a modern dynamic science of psychology. It was only when starting my PhD studies under the supervision of Professor Jack James, and having reviewed my first few lamentable attempts at writing a coherent proposal for a doctoral research programme on cognitive attentional biases in psychological disorders, that Jack directed me to a theory that he wished to find more about and that I hadn’t heard of previously, Relational Frame Theory (RFT), through his connections with Dermot Barnes-Holmes. A new Psychology Department had just been established at Maynooth University in Ireland, and at the time just had two permanent members of staff, Dermot himself and Bryan Roche. Thus, I went to visit Dermot and Bryan in person to find out more about RFT and behaviour analysis in general. They were incredibly supportive, and it just felt like a very exciting time. Bryan Roche became easily the most important influence in my career and I was lucky enough that he agreed to come on board as co-supervisor. There were very impressive PhD researchers at Maynooth around that time in Ian Stewart and Denis O’Hora, and later Robert Whelan and Louise McHugh. I was further supported by Julian Leslie in Ulster who could well be regarded as a father figure for behaviour analysis in Ireland. I have particularly fond memories of ABAI Conferences in San Francisco (2003), Boston (2004), and Chicago (2005). The work on stimulus equivalence, transformation of function, and derived relational responding more generally just seemed urgent and vital. Jack James subsequently left a remarkable legacy in Galway by hiring some of the best young talents in behavioural psychology and setting up a programme in applied behavioural analysis. I have been working in England at the University of Chichester for the last 8 years and there is remarkable resistance in the South of England. Pockets of the UK, such as Wales, are much more open to the scientific merits and power of behaviour analysis, but for the majority of the UK it is a real challenge to promote and advocate a behavioural science of psychology. The irony is not lost on me as the contingencies of reinforcement in the UK as determined by the government research allocation of funding to universities is highly focused on research ‘impact’, and as a functional psychology, behaviour analysis is primed to deliver impact by positively effecting people’s lives.

My research covers a number of different strands, initially with a strong focus on the role of stimulus function in stimulus equivalence class formation and the implications that might have for our understanding of the development and subsequent treatment of clinical anxiety, but for the purpose of this blog I will just highlight some of the current exciting developments. I will not go in to depth on what RFT is, as that has already been elucidated more clearly than I ever could in previous blogs by Ian Stewart and Siri Ming, but I think it is important to focus on the simplicity and elegance of the fact that derived relational responding is an operant, it is behaviour that is initially learned and can be shaped and developed. We believe that derived relational responding (in some publications you will see the phrase ‘arbitrarily applicable relational responding’) and the transformation of stimulus functions is at the core of human language and cognition. Our capacity to derive relations among stimuli is driven by our desire and need to communicate with each other and make sense of the world around us. One of my favorite papers of all time remains Skinner’s (1989) ‘The Origins of Cognitive Thought’ and in it you will find some of the most beautiful examples of transformation of stimulus functions ever committed to print. I think it is important to also acknowledge that RFT is not a ‘theory of everything’. We do not claim to know all the answers by any stretch of the imagination, and there are certainly other aspects that could enrich the theory such as the more recent considerations of evolutionary science and neuroscience, but do propose that the relational operant is simply a very useful concept in helping understand much of what people do. This simplicity and drive for prediction of and influence over psychological and behavioural events might appeal to those, like me, who have ever been frustrated by teaching, for example, a semester of Social Psychology to undergraduates with its umpteen different theories for every little phenomenon or empirical effect.

The first strand of research that we are really enthused about is the continued development and application of the Function Acquisition Speed Test (FAST). The FAST is rooted firmly in behaviour analysis and RFT principles in that stimuli acquire functions over time. These stimulus functions may help identify a person’s individual learning history. By empirically examining differences in rates of stimulus function acquisition in what might be termed ‘consistent’ or ‘inconsistent’ contexts with a person’s developmental history, we can potentially highlight stimulus relations that are firmly established and might be difficult to alter in say a therapeutic intervention such as talk therapy. The FAST could be considered a behavioural test of implicit cognition or attitudes (i.e., implicit behaviour) but without the procedural and scoring concerns of popular tests of social cognition such as the Implicit Association Test (IAT). The FAST has been developed over a number of years of careful experimentation and empirical refinement with colleagues such as Bryan Roche, Maria Ruiz (RIP), Anthony O’Reilly, Amanda Gavin, and Aoife Cartwright, but has been taken to a new level by the very talented Jamie Cummins. We are very proud of our recent publication of the FAST in the Journal of the Experimental Analysis of Behaviour, ‘The relationship between differential stimulus relatedness and implicit test effect sizes’ (Cummins, Roche, Tyndall, & Cartwright, 2018), and believe this to be a significant milestone in RFT’s history as JEAB remains the flagship journal for our field. We have numerous FAST-based papers under review or in preparation in a variety of basic empirical and applied contexts from work carried out in my Chichester lab and other labs. We believe the FAST has potential for utility for behavioural, clinical, social, and health researchers and practitioners and, while receiving the JEAB seal of approval has given us greater confidence over the scientific soundness of our methodology, we would encourage others to adapt the parameters of the paradigm with experimentation as true behavioural scientists should as it will naturally adapt and evolve over time. As such, we have made the FAST open access for people to use in their own labs or practice.

The second strand I will discuss is the, quite simply, remarkable series of findings from the Strengthening Mental Abilities through Relational Training (SMART) program. SMART is an online program established by Bryan Roche and Sarah Cassidy at Maynooth University. It is based on the premise that behaviours that are considered ‘intelligent’ are instances of derived relational responding, or relational frames. Standardized intelligence tests (e.g., Weschsler IQ scales) and tests of cognitive ability or aptitude are quite similar in that items generally require participants to observe or derive relations among stimuli (e.g., given a set of matrix images, choose the appropriate one to complete a particular sequence or pattern; Black is to White as Night is to ’blank’). In mainstream psychology, it is conventional ‘wisdom’ that IQ or Cognitive Ability is generally something fixed or static and something that cannot be changed. It is often viewed as an inborn and invariant characteristic of a person, and a construct that strongly predicts educational achievement and future life success. As behaviour analysts, we believe that intelligence is behaviour or that intelligent behaviour can be viewed as situated acts in context. Thus, like all behaviours, we believe that intelligence can be reinforced, shaped, and altered through interactions with the environment. RFT would propose that those who score high on intelligence tests already have a very well developed relational operant repertoire, or in other words, highly fluent in derived relational responding. Moreover, those who score low on such IQ tests have a derived relational responding skillset that can be enhanced or improved. The SMART programme, based on RFT principles, trains participants with SAME/DIFFERENT and MORE/LESS relations over a series of 55 or 70 stages (see RaiseYourIQ.com). In one study, training with just this very limited set of relations resulted in a huge increase in IQ score from baseline to post-test of an average of 23 points (Cassidy et al., 2016). Given that 15 points represents 1 standard deviation on the Wechsler scales, this is a profoundly significant finding, especially as IQ score has been shown over and over again in the psychometric literature to be the best predictor of academic success. To put this finding in context, the largest increases typically seen in the cognitive literature with interventions based on attention and working memory (such as the N-Back task) is just 2-3 points on average (Au et al., 2014). We have recently replicated this finding in an Irish school in a paper just published in the Journal of Behavioural Education ‘Can SMART training really increase intelligence? A replication study’ (Colbert, Tyndall, Roche, & Cassidy, 2018). Here in my Chichester lab, under the joint supervision of Antonina Pereira, Shane McLoughlin is a PhD candidate who has conducted the largest randomised control trials of the SMART programme to date in primary and secondary (i.e., high school) school settings. Some of these studies include step-wedge designs, all include active-control groups (e.g., computer coding), and some also include other measures of motivation, personality, and student perceptions of themselves as learners. The ‘SMART Effect’ remains strong throughout these data sets and numerous papers are in preparation. Armed with this data, Shane is taking the brave step of presenting our work at the International Society for Intelligence Research (ISIR) conference in Edinburgh in July 2018. Considering that the majority in this society are firm believers in intelligence as a fixed genetic trait, we are very pleased to get a presentation with ‘operant’ in the Title accepted into the program! We believe this line of research represents one of the big success stories of behaviour analysis and RFT, in that it has the potential to make really significant and valued differences to children’s lives. Intriguingly, Nanni Presti’s lab in Enna in Sicily has conducted fascinating research with the SMART program in an elderly population with Dementia/Alzheimer’s and has found startling findings. This is just the tip of the iceberg as SMART trains just a few relational frames, future researchers could examine effects of a much wider range of relations (e.g., deicitic, hierarchical, temporal).

One final branch of work is based on the psychological flexibility model of Acceptance and Commitment Therapy (ACT) along with my colleague and former PhD student, Daniel Waldeck, and collaborators such as Luca Pancani, Paolo Riva, Robert Whelan, and David Dawson. This has included a number of publications on psychological flexibility, experiential avoidance, and cognitive fusion in ostracism, with a number of other papers on tests of the psychological flexibility model itself. All of this work is anchored by a focus on RFT as the theory that underpins ACT.

There was a time around ABAI in Seattle in 2012 where it felt to me that RFT had plateaued somewhat. I enjoyed the conference as I always do and both my symposia presentations were fruitful, but, unlike Chicago ABAI 2008 and Phoenix ABAI 2009, there seemed little energy and little empirical work being presented on higher-order cognition such as analogy, metaphor, and categorization. However, despite the fact that RFT typically has a small (albeit very focused and determined) number of researchers in a limited number of labs around the world at any given time, I have renewed vigour and hope for the future of RFT now with some inspiring work from the likes of William Perez, Jamie Cummins, Joao De Almeida, Alice Hoon, Simon Dymond, Dylan Colbert, Louise McHugh, Ian Stewart, Teresa Mulhern, Francisco Ruiz, and Carmen Luciano among others. In saying that, it is important for us to remain focused on what RFT is and what it is not. This difference in understanding was clearly highlighted in a recent response by Dymond and May (2018) in The Psychological Record to a citation analysis of RFT by O’Connor et al. (2018) in the Journal of Contextual Behavioural Science, in that it is clearly tempting to attribute too much work to an RFT inspired approach. RFT is evolving too, of course, with a number of potentially interesting changes of direction emerging. In conclusion, I think RFT needs ABAI and ABAI needs RFT, they are not incompatible bedfellows at all. As Ian Stewart put it, behaviour analysis focuses on the operant, RFT focuses on the relational operant, and I think they are better and stronger together.

Dr. Ian Tyndall

Functional Behavioural Science Laboratory, Department of Psychology, University of Chichester, UK