Hello Everyone,
Please read the the Eichorn et al. (2016) article on the role that working memory might play in speech fluency and write a discussion board post by 9PM Monday, September 26th. I additionally included a brief supplemental article on Baddeley's conception of working memory because I thought Naomi's article assumed a lot about the reader's knowledge of the construct. So, although we won't discuss Baddeley directly, do read this too if you want to know more about working memory, perhaps reading it first to serve as a primer for Naomi's article.
Naomi will present on this research so your discussion posts can lead you to ask questions in the class portion, but also in directing questions towards Naomi in the public presentation.
Remember that a good portion of the grade for this class is based on your participation, not just in terms of posting a response to the readings (10%), but also in terms of the in-person discussion (15%). Whereas some students are more willing to openly discuss the topics, still others are relatively silent in the course. Please consider it a safe space to discuss your ideas openly.
I am interested to discuss the ideas with you at 3PM on Wednesday, September 28th.
Hope you are all well!
Best,
Dr. Braasch
This week’s readings were not only interesting in regards to research, but a good refresher on underlying cognitive topics such as working memory (WM). Since better WM has been shown to be associated with better overall cognition in many respects, it seems appropriate to use it as a tool to examine where deficits occur. Although, the results of the study make me question whether these measurements were appropriate for studying the differences between those who stutter and those who do not. However, I am curious to hear Naomi’s discussion, because I am not too familiar to this area and am not quite sure what to make of their findings, though for the discussion will give my best effort.
ReplyDeleteTo begin, the baseline speech measures showed no difference between those who stutter and those who did not. Again, my naiveté may be showing, but it seems to me as though there should be a baseline difference between these two groups if one has a speech impairment. To this extent, I think it would be interesting to find another way to measure baseline speech fluency that is more sensitive to detect differences between these populations for use in replication.
Additionally, to my knowledge there were no WM differences between the two groups. This, to me, provides evidence against the claim that “stuttering arises from excessive speech-monitoring activity and that limiting attentional or working-memory (WM) resources available for such monitoring may improve fluency.” In other words, while they did show that dual-processing enhanced speech fluency, it was no different for those who stutter compared to those who do not. As explained in the implications, more research needs to be done to explore if this dual-processing would actually have any sustained positive effects for those who stutter.
In regards to replication, I definitely think it would be worthwhile to replicate this study addressing the issues described in the limitations section, such as increasing WM load in the tasks, using WM tasks that are comparable, controlling time for speech preparation, and examining other aspects of linguistic productivity, and possibly using a larger sample. Replicating with these changes may make measures more sensitive to group differences and increase reliability and accuracy of the study.
I am writing this right after reading the papers after a week of conferences. So I apologize for brevity and/or inanity.
ReplyDeleteI am not very familiar with memory systems, so I really appreciated how the first article gave background before jumping into the new material it brings to the field. It makes sense to me that the complexities of memory would be more than a single system or store. I do wish it had gone in to more detail about the tasks used to simulate short-term memory deficits, since I am not sure how they were performed.
The second paper was pretty neat. I’ve participated in working memory research in the past I wonder if the reason that allocation of WM increases speech fluency is that speech is an almost instinctual process by adulthood and that not thinking about it makes it somewhat easier. I think the fact that there was improvement for both stutterers and nonstutterers and no significant difference between the two supports this thought. I guess it’s really just my attempt at putting the CAH into laymen’s terms.
This comment has been removed by the author.
ReplyDeleteThis article is very interesting, because before I read this article, I never thought that sometimes less cognitive control can lead to a decrease in disfluency for adults who stutter, I used to thought the opposite. This article is also very useful because I think it is a very good example when considering how to write method section in a research paper. Overall, the method section is very detailed and well organized, especially the detailed description of the procedure of four sets of tasks, and how the authors processed the data. I think one of the benefits of providing an explicit method section is for future replication. In the literature review section, the authors believe that one of the reasons why dual-task studies on this topic are inconclusive is because of critical methodological differences, such as not distinguish between different types of disfluency or fail to specify the cognitive resource targeted by secondary tasks. Therefore, an explicit methodology section is very crucial for future replication.
DeleteThis may be a silly question because of my lack of knowledge in this topic: one thing that I am curious about in this article is that since the authors mentioned disfluencies as a result of covert repairs, hypervigilance, and explicit attention, my question is that how can researchers distinguish a person who stutters is because of hypervigilance not explicit attention. In addition, as regard to the implication, because the findings show that engaging working memory while speech was produced caused a significant reduction in atypical disfluency for adults who stutter, and authors suggest that future research can focus on determining whether and how dual-task methods can supplement traditional stuttering therapies. So I wonder whether this method can also be applied to children.
I really enjoyed these two articles. I had never given much thought to what causes stuttering, but the longer of the two articles did a great job explaining possible causes and treatments. It makes sense that distracting WM during speech for someone who stutters may allow them to think less about their speech and therefore increase fluency. I think this notion is actually a very familiar concept for most- when you overthink things, you tend to mess them up! That's an experience I am sure we've all had so relating that to stuttering was simple and straightforward.
ReplyDeleteI found it interesting that the two groups, people who stutter and those who do not, showed no differences in their ability to handle dual-tasks simultaneously. What makes speech different for those who stutter? It seems odd to me that stutters remain relatively stable through adulthood, even though speech becomes more natural and easier as we age.
For replication, I would really like to see this study done on children with stutters, as well as older adults. If working memory itself evolves through the lifespan, we should see different results for different age groups of participants. I also wonder how the one participant who had not received any therapy for their stutter differed from the group. Does therapy for stutters involve tasks like this? The article mentioned that therapy teaches stutterers how to focus on certain parts of speech which seems entirely counterproductive according to the results of the study.
Thank you for including the short article on working memory! I had a vague idea of what it is and the article helped answer some of my specific questions! It makes sense that WM is made up of multiple components. It’s clear that WM can improve cognition, and I was interested in reading about how stuttering might impact that influence. I was a little surprised that there were no differences between the two groups in baseline speech measures. I thought there would definitely be a difference, and maybe even a large one. I was also surprised that incorporating WM into a treatment plan for stuttering decreases the frequency of stuttering.
ReplyDeleteIn terms of replication, I would love to conceptually replicate using children. This was mentioned in the “Future Research” section of the article. I would be interested in seeing how they react to the WM task, since they haven’t been stuttering as long as the adults in the present study. I am curious as to whether or not the amount of time spent with a stutter impacts the results. I would also like to see this done in other languages, because I don’t know if WM and stuttering are impacted by different types of language (like more harsh languages, or languages that have more flow). Another possible way to replicate the study involves investigating how emotion regulation and attention regulation could work together to impact WM, and thus reduction of a stutter.
This comment has been removed by the author.
ReplyDeleteThe claim that implicit system involved in motor function is not available to WM is suspect. I think the authors are correct that this kind of implicit content is not easily expressed verbally - I would call it non-propositional - but that does not mean that it is not available to WM. A problem arises, of course, if we consider the contents of WM to be propositional or, as in practice we often do, we express contents of WM verbally. But, it seems implicit motor systems do need a kind of working, functional, short-term memory (albeit non-propositional). Where your arm is now is a function of where it was and where it can and will be is a function of both. Motor intentionality towards past motor states is necessary for the motor actions we take; I'm not sure what to call this if not "WM." Motor "memory" becomes all the more important when we consider the notion of skill and, in particular, the development and cultivation of skill.
ReplyDeleteIt also seems unsurprising that "external" focus improves motor functionality, since - representations or not - the things we do involve the world external to us; the ball I catch is not inside my skull, even if some philosophers and scientists may want to claim it is represented there.
I have a concern with defining automaticity, in part, as a task's not being affected by other ongoing activity. It seems like it will then be trivial that, given some task is affected by other activity, it is not automatic. The concern, then, is that very, very few tasks are actually automatic - I'm not even entirely sure that the beating of your heart would count, as certain mental activity can in fact disrupt it. If the claim is narrowed to regard activities indexed in time, this will affect the generalizability of automaticity claims, and perhaps the interpretation of experimental data as well. There is also an odd claim that I am having trouble resolving with the definition of automaticity: dual-task conditions enhanced speech fluency. However, this is a case of other mental activity positively affecting disfluency... By definition, then, atypical errors are not automatic, so it is not a particularly interesting claim that they represent a break down in automaticity. I think this objection on my part may be a bit smoke and mirrors, but I am meaning to point at what I take to be an underlying and more general issue of allowing your definitions to affect conceptualization and analysis in unintended and perhaps undesirable ways.
I see the pool of participants was drawn from a wider source than a psych 101 class, but I wonder if individual demographics were recorded or presented. It is possible that, though flyers were distributed "widely" many persons who participated were still students, for example, and that may be relevant to replicating the experiment as well as interpreting data.
I'm not sure I could speak "continuously" for 60 seconds on the topic "What's in a name?" The complexity of a topic shouldn't be conflated with disfluency. Generally, content ought not be confused with form. This might be "normalized" in a variety of ways, but it seems like a needless wrench in the mix.
This article portrayed a number of the difficulties with replication. For instance, in the literature review, the authors mentioned that some of the findings related to stuttering are inconsistent in previous research, with negative, neutral, and positive relationships all being observed. In these instances, particularly when evidence is almost equally distributed, I wonder how one best chooses the evidence upon which to build one's hypotheses. In addition, the article highlights the added difficulty of having multiple theories/explanations that could conceivably apply to a given phenomenon. In the instance provided here, all three hypothesized explanations for stuttering had previous support, albeit with inconsistencies, and provided distinct mechanisms underlying the same outcomes. Whereas a number of researchers might only draw from one theory within a single study, I found it interesting that the authors of the present article addressed all three distinct hypotheses simultaneously. Further, I believe this provided a useful glimpse at how one might modify different aspects of a study to tap into nuanced elements of certain topic. I thought that the authors did a great job of using very specific manipulations to address unique differences among the competing theories/hypotheses. By doing this, the authors managed to provide replication results for all three, while also distinguishing the CAH as the explanation that best fit the entirety of the results in the present study. However, this did raise some questions for me. For instance, while the CAH seemed different enough from the CRH and VCH to allow for one to evaluate which underlying assumptions were being best replicated, it seemed more difficult to make these evaluations between the CRH and VCH themselves, as these overarching hypotheses were much more similar. In these instances, where there are competing explanations with differential and sometimes inversely predicted outcomes, I wonder when one might consider failed replication as the result of methodological error or an unaccounted factor and when one might consider it as potential support for the conflicting theoretical assumptions.
ReplyDeleteThe concept of working memory, as well as distinctions regarding what it is (e.g. Engle’s view of WM as more an ability to inhibit non-goal-related stimuli, versus Baddely’s original view), what its parts are, and how it is measured, is fairly central to cognitive psychology and discussions of processing. As such, it does not seem as novel an idea for discussion as the second article.
ReplyDeleteIn Eichorn et al., various hypotheses regarding the cause of stuttering were presented, all of which seemed to provide a fairly sound explanation, and all of which seemed to share equally ambiguous support. I have never personally read much on stuttering, so it was interesting to see how this was tied into other research to propose the mechanisms by which it occurred. For instance, I had previously heard discussions of the detrimental effects of deliberative processing of complex motor movements in the context of darts through some popular source (e.g. documentaries), but had never considered its possible application to speech which, in essence, is just an incredibly complex sequence of motor movements which we connect with ideas.
In various discussions and investigations regarding the benefits and detriments of more deliberative processing, it has been shown that, although this processing is incredible for math, logic, or complex reasoning tasks, it is simply too slow and inefficient for complex motor behaviors which are better learned as a series of actions than if they are learned through study (e.g. attempting to learn how to ride a bike by reading). It is also interesting to note that the hypervigilance models have shown increased activity in the Anterior Cingulate Cortex (ACC) among people who stutter, as this region is frequently associated with conflict detection in other contexts, as well as the recruitment of more deliberative processing resources. For this reason, the constrained action hypothesis (CAH) seemed to make a great deal of sense to me.
There was one self-induced concern I had, which maybe others were confused by, as well. At first, the speech fluency results seemed a bit strange. AWS produced significantly more disfluencies (particularly, atypical disfluencies) than AWNS, which seemed to be in line with predictions, but the interaction of disfluency and speaker type within the dual-task procedure was not significant. HOWEVER, it then occurred to me that the reason there was no interaction is that the AWS no longer produced a significantly different number of disfluencies than AWNS under the dual-task condition, suggesting that this procedure had essentially eliminated the original difference. I would like to state the author should have made this more explicit but, in reality, I believe this was my own error.
In light of this, then, it seems that the CAH does provide a good fit for stuttering and that stuttering may best be conceptualized as a tendency to attempt to apply more cognitive resources to a process which simply does not lend itself well to this more reasoned and deliberative style of information processing. Additionally, although this explanation explains what is occurring, it does not necessarily mean that other hypotheses regarding the development of this behavior are invalid. For instance, the vicious circle hypothesis (VCH) suggested that there may be a “hypervigilant alarm mechanism” (see p. 2). This fits very neatly with the observation that there is increased ACC activity which is frequently described as a conflict monitoring system (e.g., Botvinik, Cohen, & Carter, 2004) and is associated with the allocation of more resources to the task. In this sense, it may be that conflicts are detected too easily in regards to speech, thereby allocating more resources, which then decreases performance (as suggested by CAH). In this way, a slight modification of the VCH is not only not contradictory to the CAH, but may help elaborate how this disposition to more explicit processing develops.
I thought that this article was rather interesting, especially with its relation and applicability to replication. I have personally noticed automaticity influencing performance in something like playing sports, and can anecdotally relate that phenomenon easily. I can also think of examples of when people claim that they feel like they can speak a foreign language easier when they consume alcohol.
ReplyDeleteNot really having much background about stuttering or speech disorders, it didn’t really occur to me to think that they would have differences cognizing, although the article certainly makes a fascinating case about why they might would think that it would improve performance. This idea is demonstrated by the fact that engaging working memory resources while speaking improved enhanced the fluency. This idea of automaticity in sports and other domains was appealingly conceptually reproduced to show apparent increases in fluency, which related to the overall theoretical models that they used with respect to understanding the cognitive process that occurs and is responsible for stuttering.
I’m particularly interested in the potential for clinical interventions, especially with children. Fluency with regards to speaking other languages might also be an area of interest, as I imagine that there is an element of “overthinking” and a lack of automaticity that serves as a barrier when people attempt to formulate sentences in nonnative languages. I imagine that the a similar study utilizing a working memory task and evaluating fluency in a nonnative language would yield similar results to this study, and that you would see an improvement in fluency.
One thing I thought was interesting that was mentioned in the Baddeley paper was how working memory capacity is a good predictor of how well new members of the Air Force will learn certain things and that it is an even better than some standard “scholastic” measures. It made me think about what would happen if measures of working memory capacity were used in the admission process for college or graduate school. I feel like this study has already been done, but I would be curious to see what kind of predictive power it has over success in graduate school. I would also be curious to see how it compares to the commonly used criteria of undergraduate GPA and GRE scores. I do not think that it is realistic to use a certain threshold level of working memory capacity as admissions criteria though because there are ways to compensate for lower working memory capacity.
ReplyDeleteIn the Eichorn et al. article, the concept of low working memory capacity being helpful in the case of speech fluency is completely novel to me. When I read the abstract, I actually thought that they wrote it backwards by accident for a second until I stopped to think about it. Something that confused me about this article though, was how in the introduction they talk about previous studies that used dual tasks and how their results are inconsistent with each other. Yet the current study found that working memory load lead to increased fluency. They say that the previous studies varied greatly in the types of tasks used in them (e.g., amount of interference, automaticity, and time overlap), which is why their results might be so inconsistent with each other. Then can we only conclude that it is the specific combination of tasks used in the current study that lead to the current results? I think it will be important for future research to try to systematically vary the tasks to see how far the findings generalize.
I found this week’s articles interesting and very novel. I am not familiar with this type of research so a lot of the material was very new. Past research has produced mixed results when examining CRH. CRH is a model of speech production in which disrupted and disfluent speech are produced when editing occurs after articulation has already started. Which is an encoding problem. Dual task studies have also been inconclusive. Disfluencies have decreased in conjunction with distractions, but this finding has not been found consistently across studies and the opposite has also been found. I think it is safe to say that consistent replication in this field of study is very difficult. I did find it interesting that Eichorn et al. looked at disfluencies as a result of explicit attention. It would seem to me that with all of the conflicting past literature, re-framing the subject from a new perspective could be hugely beneficial. I also found it interesting that they did not find differences in working memory between groups. I think there would have been differences between groups if decreased monitoring improved fluency. I think that for future replication, looking at the speech of children may be enlightening. Depending on the age of the child, it could be beneficial when determining why the stutter is happening. For example, I would guess a very young child is less likely to exhibit a stutter because of explicit attention. However, that might be wildly incorrect as I have no background in this literature.
ReplyDeleteFirst I would like to state the obvious by saying I think these articles were great, as I personally enjoy studying memory, more specifically working memory. I found myself completely intrigued by the nature of the Eichorn article. I find language very interesting (but have very little knowledge on the subject) but had never given much thought to the difficulty in studying speech impairments such as stuttering. I feel the authors did a great job in highlighting that past research has yielded inconsistent results in the introduction and carrying on to discuss this again at the end of the study. For example, in the discussion section regarding dual-tasks effects on overt measures, their findings varied significantly from past research. Their suggested explanation for this difference lies in what is considered an error among different researchers as well as in difference of measures. They bring up how difficult studying covert speech errors is specifically, and how they technically cannot be reliably measured or defined. I think they were very ambitious in tackling something as complex as this and this is a good example of how to face replication challenges when past research has not be consistent. I feel the difficulty in replicating this subject matter is made very clear as previously stated, and I wish I had more insight into understanding how to make it less difficult.
ReplyDeleteAside from my admiration in their replication of past work, I did find the results from the study to be interesting. While their original hypothesis was supported, the two groups did not differ in their ability to handle dual tasks simultaneously. I am having a hard time understanding this lack of difference and wonder if it has something to do with testing only adults? I think replicating this with children would be interesting, as their language development would not be nearly as concrete as adults. I am also curious, as elementary as it sounds, would varying levels of stuttering severity be another way to address this subject? I did a quick google scholar search and read in an abstract that there is much debate and inconsistency in research regarding stuttering in general. It seems that certain aspects of language research are faced with the challenge of finding valid and reliable measures for their constructs, something I found interesting.
This article successfully provides experimental evidence which showed that excessive cognitive processing, especially working memory, which is occurring simultaneously with speech production causes speech disfluency. In other words, a specific optimal level of cognitive control is required for the effective performance, especially in speech production.
ReplyDeleteThe most interesting part of this article is that this speech disfluency showed in both AWS and AWNS groups, regardless of participants' stuttering status. Based on this results, I am curious that Whether or not I can replicate this research with normal healthy bilingual speaker groups. In bilingual research, there are lots of studies investigating whether first and second language interrupt or enhance one another’s production. In some studies, they provide evidence which shows that first language and second language enhance each other with sharing semantic networks, since more connections mean stronger function in networks. On the other hand, there are also plentiful studies indicating that first language interrupts second language production. If I could put first language processing as secondary dual task, can I design a research examining how second language production fluency would be influenced under first language processing condition? For example, measure behavioral results and brain activity when bilingual people studying second language stuff while listening first language music…
I know it’s unorganized, but as this article showed that controlling excessive processing enhances speech production, if bilingual person control their first language, maybe there is a chance to improve second language production.
I found both of this week article is very interesting. In the Eichorn piece it is found that changes in fluency were accompanied by reductions in speaking rate but not by corresponding changes in overt errors. Moreover, WM contributes to disfluencies regardless of stuttering status and that engaging WM resources while speaking enhances fluencies. While I did research on cognition I did not find that much about dual-task conditions enhanced speech fluency. May be I did not look for that.
ReplyDeleteIn the literature part of this article previous research shows that CRH has produced mix results. Phonological skills in PWS and persons who do not stutter at least subtle phonological weakness among individuals who stutter compared with individual who do not on the basis of analysis of spoken output.
Dual task studies that attempted to suppress speech monitoring have similarly inclusive. The preset study examined two related predictions of the CRH, 1) error signals generated by speech monitor are received by WM, 2) tax WM were expected to affect monitoring integrity.
Results also shows that a main effect of domain with verbal task being more difficult than special task overall but no domain task interaction. Corroborated expected differences in a dual task effects depending on disfluency type with significance reduction in atypical forms of disfluency type.
In the speaking task speech preparation time was not controlled and this makes the topics more difficult to determine weather language formulation and conceptual planning differed between groups or contributed to the results.
There were a few items I found interesting in the results of this study. First, the lack of a significant difference between the baseline measures. I'm sure this largely due to misconceptions and a lack of history on my part, but I expected to see differences in the baseline measure. The fact that no differences existed and the results were able to be collapsed into a single measure just seems somewhat unexpected for me.
ReplyDeleteSecond, the effect of dual task effects on speech fluency is interesting. Their results on overt errors suggests the predictions of CRH to be incorrect. CRH seems to have mixed results.
Just as a note as to the sample of this study, while the sample was generated from a larger group than intro psych classes, perhaps some re-testing on a different group could be interesting. The mean age in the study is in the mid to late 20s. It would be interesting to see how these results hold up in children, especially given the discussion of treatment intervention. If these studies are to give us any applied knowledge, and we know treatment at a young age is important, these results need to be confirmed in that population.
Having some background in linguistics but no background in speech pathology, I found this article particularly interesting. Growing up, my best friend in grade school had a stuttering problem but of course, as a 10 year old, I never thought to myself “I wonder what his working memory capacity is and if he’s engaging in covert repairs?”. Though stuttering is a widespread problem, especially amongst males, I was unaware of the major theories put forth to explain the phenomena. I found all three of the major theories to be well grounded with CRH having the clearest ties to working memory within the Baddeley model—with particular emphasis on the phonological loop and the speech planning and rehearsal that would be carried out there. If a deficit in working memory was observed, then I felt that it would make sense that AWS would have a far more difficult time repairing speech errors within the phonological loop before the error laden speech is articulated. However, the main finding of the study spoke to a type of overload that needs to occur within WM in AWS and it was this overload that more or less “distracted” AWS and helped to enhance fluency.
ReplyDeleteAs far as replication goes, I believe that this line of research would benefit from both direct and conceptual replication—mainly due to the clinical nature and application of this type of research. If research like this can be directly replicated, then speech pathologists can focus more on actively distracting PWS from their speech in order to help repair it. In developmental psychology, parents are often encouraged to engage in “active distraction” in order to guide children’s awareness away from unpleasant stimuli. Perhaps something similar to this could be adopted if direct replication is indeed achieved. I also believe that conceptual replication would be paramount here, especially in early intervention practice as the future directions of the article speaks to. If speech pathologists are able to identify and treat PWS with this WM overload and distraction technique early enough, it may be able to lessen the allocation of attentional resources away from their speech disfluency and instead toward fluency enhancement throughout the lifespan.