Executive function, working memory and speech-in-noise recognition – Comparing a non-semantic black and white version of the Trail Making Test to the original Trail Making Test

(1)

Linköping University | Department of Computer and Information Science Bachelor thesis, 18 HP | Cognitive Science Spring term 2017 | LIU-IDA/KOGVET-G--17/033--SE

Executive function, working memory and

speech-in-noise recognition – Comparing a

non-semantic black and white version of the

Trail Making Test to the original Trail Making

Test

Marc Friberg

Supervisor: Rachel Ellis Examiner: Björn Lyxell

(2)

Upphovsrätt

Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare – under 25 år från publiceringsdatum under förutsättning att inga extraordinära omständigheter uppstår.

Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner, skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för ickekommersiell forskning och för undervisning. Överföring av upphovsrätten vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av dokumentet kräver upphovsmannens medgivande. För att garantera äktheten, säkerheten och tillgängligheten finns lösningar av teknisk och administrativ art.

Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i den omfattning som god sed kräver vid användning av dokumentet på ovan beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan form eller i sådant sammanhang som är kränkande för upphovsmannens litterära eller konstnärliga anseende eller egenart.

För ytterligare information om Linköping University Electronic Press se förlagets hemsida http://www.ep.liu.se/.

Copyright

The publishers will keep this document online on the Internet – or its possible replacement – for a period of 25 years starting from the date of publication barring exceptional circumstances.

The online availability of the document implies permanent permission for anyone to read, to download, or to print out single copies for his/hers own use and to use it unchanged for non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility.

According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement.

For additional information about the Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its www home page: http://www.ep.liu.se/.

(3)

iii

Abstract

In this thesis, the relationship between cognition and speech-in-noise recognition, in normally-hearing Swedish students, is examined. The Trail Making Test, hypothesized to measure a wide range of cognitive functions, including executive function and working memory, has been criticized for being a culturally biased measure, hence the need for a culturally unbiased version. A between-group experiment was conducted in which a non-semantically dependent version of the Trail Making Test was compared to the original Trail Making Test in order to test for psychometric equivalence. A total of 21 young normally-hearing Swedish students were given three tests: TMT or TMT (non-semantically dependent version), a Swedish Reading Span Task and a Swedish speech-in-noise recognition task. The B parts of the two Trail Making Test versions differed significantly and both were moderately to highly correlated to speech-in-noise and reading span performance. The results indicates that the original Trail Making Test is a more plausible index for executive function and strengthens the relationship between executive function, working memory and speech-in-noise recognition.

Keywords: cognition, trail making test, speech in noise recognition, reading span task, working memory, executive function

(4)

(5)

v

Acknowledgements

I would like to thank my supervisor Rachel Ellis for all the valuable support, tips and discussions throughout this thesis work. I have learned a lot. I would also like to thank Björn Lyxell and fellow classmates for continuous feedback during the seminars; and everyone else involved during this project. Last but not least, I would like to thank Linnea Fornander for bringing out the best in me.

(6)

(7)

vii

Introduction

Communication is essential for humans. And one of the most valuable and essential abilities we have in communication is the ability to listen to each other. Listening is not just the passive process of receiving sounds from the outside, as in hearing, but also processing the sound internally in the brain with attention and intention using your cognitive abilities (Kiessling et al., 2003). In the emerging field of cognitive hearing science, the role of cognition in speech recognition in people with hearing impairment is widely acknowledged (Arlinger et al., 2009, Pichora-Fuller, 2007). Focus has shifted from tackling hearing deficits from an engineering perspective or researching how single word stimuli is perceived to understanding how speech recognition functions in a more naturalistic everyday environment (e.g., Arlinger et al., (2009)).

Thus, the ability to listen and perceive someone speaking is not independently determined by the acoustical properties of the speech, but also on the surrounding noise and an individual’s cognitive abilities. If, for example, you and I were to be talking and a third-party person were to be speaking in the background at the same time, you would be distracted to a certain degree, and the degree that you would be distracted to is believed to partially be determined by your working memory capacity (Stenbäck, Hällgren, & Larsby, 2016).

Recent findings have showed that the Trail Making Test is a suitable tool for predicting speech-in-noise recognition (Ellis et al., 2015; Ellis & Munro, 2013, 2015). However, the Trail Making Test is in some cases not usable in populations of non-Latin alphabet users (see for example H. J. Kim, Baek, & Kim, 2014; Maj et al., 1994). The purpose of this thesis is to test whether a non-semantically dependent version of the Trail Making Test is psychometrically equivalent to the original Trail Making Test and examine the relationship between Trail Making performance and working memory capacity and how it influences the ability to perceive speech in a noisy context.

(10)

2

2. Theory

In this section the theoretical framework that this work is built upon will be explained.

2.1 Executive functions

Executive function (EF), sometimes also referred to as executive control or cognitive control, is a cluster term for three core top-down cognitive functions: inhibition, working memory and cognitive flexibility (Diamond, 2013; Lehto, Juujärvi, Kooistra, & Pulkkinen, 2003; Miyake et al., 2000). Although the three functions might be considered completely separate, each function will be briefly explained and its relation to the other functions will be described.

Inhibition, as will be denoted inhibitory control from now on in this text, refers to the ability to control ones’ attention, behavior and thoughts. Without this ability, we would fall victim under our impulses unable to steer our behavior and achieve our goals. Sometimes we are unable to control our reaction to stimuli. Important to note is that inhibitory control is not to be considered to be one single controlling mechanism, but that there are several types of inhibition (Diamond, 2013). At a perceptual level for example, we have what Diamond (2013) calls interference control, when our attention is involuntary the stimuli is called exogenous and our attention is automatic since it is stimuli-driven (Theeuwes, 1991). But often we can choose to ignore a stimulus, in order to complete a certain task or goal determined beforehand (Posner & DiGirolamo, 1998). Diamond (2013) acknowledges another aspect of interference control which is cognitive inhibition. The ability to suppress previously stored information, or mental representations, such as unwanted memories (Andersson & Levy, 2009) or resisting proactive interference (Ellis & Rönnberg, 2015). On a more behavioral level, Diamond (2013) refers to inhibitory control as self-control, which is the ability to for example resisting temptations, such as eating chocolate when on a diet. Very often self-control involves the process of resisting something that one part of you might want to do (Hofmann et al., 2009) but sometimes there are no competing stimuli, such as when you resist to say something you maybe not are supposed to say, which is typically measured by a go/no-go task (see for example Gomez, Ratcliff, & Perea, 2007).

The second core EF, Working memory, as will be explained more in the next section, involves holding and manipulating information, which is no longer perceptually present, in the brain (Baddeley & Hitch, 1994).Working memory allows us not only to process new incoming

(11)

3

stimuli, but also process the new stimuli while maintaining and manipulating previously stored stimuli accordingly, which in turn makes us able to connect seemingly unrelated information so that it makes sense to us (Diamond, 2013). Working memory relates to inhibitory control since in order to achieve a goal for example, you need to know what input signals to inhibit, while keeping the goal in mind. Stenbäck, Hällgren, & Larsby (2016) explains this in the context of listening as in order to focus on a target voice, you need to inhibit competing noise or speech.

The last core EF, cognitive flexibility, is in a sense dependent on the inhibition and working memory. Diamond (2013) explains cognitive flexibility as the ability to change perspective, both spatially and interpersonally, or being able to adjust to changed you point of view in a discussion or prioritize your work tasks in a changing environment (Cañas, Quesada, Antolí, & Fajardo, 2003). Typically cognitive flexibility is measured using some type of task-switching or set-shifting task such as the Wisconsin card sorting test (Milner, 1963). Experimentally during a task-switching task, there is usually two tasks to complete, and the participant is supposed to alternate between the tasks (Diamond, 2013).

2.2 Working memory

Historically, theories of the working memory are among one of the most debated and researched areas in cognitive psychology. The theoretical concept of working memory is complex and not easily pinpointed. The foundation of different working memory models explained in this article rest on decades of work in the more general research area of memory models.

In Miller’s (1956) well-cited article, he proposed some general features of how information is processed in the brain. Miller approached the brain as a system of input- and output-signals that could be measured, and he claimed that our ability to store and retrieve information was of limited capacity. Broadbent (1958) in turn proposed his model of selective attention where he, like Miller, claimed that the information processing system of humans was of limited capacity, and therefore a filter for selective processing was needed. Jumping forward in time, Atkinson and Shiffrin (1968) proposed the multi store model of memory which at the time of publishing was one of the most influential models of memory (Bower, 2000). The multi store model conceptualized memory in terms of three memory stores: a sensory store, a short-term store and a long-term store.

(12)

4

The model though was soon to be challenged as pointed out by Baddeley (2012). The linear approach in Atkinson and Shiffrin model where information was assumed to be streamlined from the sensory store into the short-term store and finally into the long-term store was criticized for being an over-simplified box model (maybe a common problem in the world of psychology). In the study of pattern recognition for example, how a stimuli is perceived is dependent on a lot of features in the stimuli itself such as the invariance in brightness, shape or position, all which cannot be explained purely linearly, but must also be explained in a hierarchical manner (Sutherland, 1968). Instead, Craik and Lockhart (1972) came up with the levels-of-processing (LOP) model, postulating that memory was not made up of three different stores, but rather that memories were encoded at (an infinite amount of) different levels, hence the name “levels-of-processing”. Baddeley and Hitch (1974) challenged the view of Atkinson and Shiffrin (1968) by showing evidence that short-term memory was more than just one single storage component and derived that there must be at least some other component that itself consisted of several components, and the suggestion was the working memory. The working memory model can be seen as an extension, rather than a replacement, of the levels-of-processing framework as it is integrating key concept from the LOP framework with new functional pieces (Sternberg, 2012).

Baddeley and Hitch’s (1974) multicomponent model initially consisted of three components: the central executive, which controls the two other systems, the phonological loop and the visuo-spatial sketchpad. Baddeley (1996) then identified four possible functions for the central executive: focus attention, divided attention, task-switching and the last function, which eventually led to an expansion of his model, interfacing with long-term memory. Baddeley (2000) then added the episodic buffer into the multicomponent model, it is assumed to act as a buffer store between the subcomponents of working memory but also as a linking device between perception and long-term memory (Baddeley, 2012).

One of the most commonly used tests of working memory span is the Reading Span Test, which was originally created by Daneman & Carpenter (1980) and revised by Rönnberg, Arlinger, Lyxell, and Kinnefors (1989). In the Reading Span Test, participants are instructed to read aloud a set of sentences and then try to recall the first or last word in every sentence. The maximum number of correctly recalled words (sometimes only including words recalled in correct serial order) is referred to as the participant’s working memory span. There are several cognitive processes involved during the task: the process of encoding visual patterns

(13)

5

when reading the sentences and accessing their semantic, syntactical and referential relations as well as storing the final words of each sentence (Daneman & Merikle, 1996). Daneman and Carpenter’s (1980) reading span task turned out to be a good predictor of reading

comprehension and has also been interpreted as an index of a domain-general working memory capacity (Turner & Engle, 1989, Conway et al., 2005).

The distinction and conjunction between working memory capacity and executive function is however not entirely clear. The term executive attention is sometimes used to denote the fact that the two terms are closely linked together and some suggestions is that individual

differences in working memory capacity is a reflection of the ability of attentional control, thus raising the question whether there is a separate attention mechanism or not (Engle, 2002; Kane, Bleckley, Conway, Kane, & Bleckley, 2001).

2.3 Working memory and its connections to speech recognition

Rönnberg et al. (2003, 2008, 2013) have, in a range of studies modelled the relationship between working memory and speech recognition based on numerous studies where several direct cognitive predictors of speech recognition have been identified: sentence completion (Lyxell & Rönnberg, 1989), context free word decoding (Lyxell & Rönnberg, 1991),

information processing speed (Rönnberg, 1990) and rhyme decision speed (Lyxell, Rönnberg, & Samuelsson, 1994), and three indirect predictors: VN130/P200 visual evoked potentials (Rönnberg et al., 1989), reading span test performance (Lyxell & Rönnberg, 1989) and general verbal ability (Lyxell & Rönnberg, 1992). All the different predictors are in many ways intertwined and related to each other but one of the most contributory predictors of speech recognition is working memory. In a review conducted by Akeroyd (2008), a sample of 20 studies examining the relation between speech-in-noise recognition and cognition were summarized and analyzed. In one of the studies, the results showed that the among aging people approximately two-thirds of the systematic variation of speech was due to high-frequency of hearing- loss because of aging, and approximately one-third was due to general decrement in cognitive functions (van Rooij & Plomp, 1990). As such, low performance on different speech recognition tasks might not be due to physical changes in the ear, as it easy to believe, but due to some kind of cognitive decline. This in turn implies that there might be cases of misdiagnosis or mistreatment where people with cognitive deficits are treated as if their hearing deficits are due to being physically ill whereas they are not and thus do not receive appropriate treatment. Thus cognitive deficits can lead to hearing loss, but as studies

(14)

6

have shown is that hearing loss, often due to age-related reasons, can cause cognitive dysfunction, such as dementia, as well (Lin et al., 2011; Lin & Albert, 2014; Uhlmann, Larson, Rees, Koepsell, & Duckert, 1989). According to Akeroyd (2008), the data points to the fact that the reading span test was amongst the most successful cognitive predictors of speech in noise recognition. Although RST did not always significantly relate to speech-in-noise recognition, it still gave an indication of its predicative power and importance.

Rönnberg et al. (2003, 2008, 2013) have suggested the Ease of Language Understanding (ELU) model as a framework to understand the importance of cognition in speech recognition. The ELU model shares some similarities with Baddeley’s (2000) working memory model as the RAMBPHO element in the ELU model nearly acts as an episodic buffer when multimodal perceptual (phonological) input is matched against previously stored information in the long-term memory (Rönnberg et al., 2013). In short, the ELU model suggests that the processing of the multimodal input works in two ways, by implicit and explicit processing. We can think of this in terms of a top-down versus a bottom-up approach. Input is first processed in the RAMBPHO element and where the signal is either matched to previously known knowledge stored in the episodic long-term store, by lexical access processing speed elements (this would represent the bottom-up part) or if no match is found however, the signal is explicitly

processed in the explicit processing loop (this would represent the top-down approach). An example of explicit processing could be you listening to someone talking but maybe you missed a word or two, then you would actively try to comprehend what the other person were saying, and this is where high-function cognitive processes are involved (Rönnberg, Rudner, Lunner, Zekveld, 2010). Thus, it is hypothesized that as processing demand raises the storage capacity declines, and as such the greater the working memory one have the easier it will be to recognize speech in a noisy environment as more resources are available for explicit

processing (Rönnberg et al., 2013). This implies that people with low working memory capacity might have a harder time trying to perceive speech in a noisy environment.

2.4 Trail Making Test

The Trail Making Test (TMT) is widely used neuropsychological screening tool, often used in larger test batteries, used for detecting neurological alternations (Reitan, 1958, 1992). The TMT is a pen and pencil test consisting of two parts, A and B. In TMT part A (TMT-A) there are 25 encircled numbers in a scattered pattern. The goal is to connect the numbers in

(15)

7

numbers and letter are scattered across the paper and the goal is to connect all the numbers and letter in ascending order while alternating between the numbers and letters. Part A and B are hypothesized to measure different cognitive abilities. Arbuthnott and Frank (2000) describes how the test initially was interpreted: TMT-A was seen as a baseline measure for motor and visual control as well as speed, whereas for TMT-B, the addition of having to switch between numbers and letters would require additional demand on executive functions. In a review by (Sánchez-Cubillo et al., 2009) they suggests that the cognitive abilities

measured by TMT-A are mainly visuoperceptual, and the abilities measured by the TMT-B involve, in addition to those measured by TMT-A, working memory and task-switching. Sánchez-Cubillo et al.:s (2009) study showed that in healthy old adults (Mage = 59.4) the B-A

measurements would be a pure indicator of executive function.

The Reading Span Task turned out to be a significant predictor of speech recognition, as mentioned in the earlier section, and the results of, for example, Arbuthnott and Frank (2000) and Sánchez-cubillo et al. (2009) opened up the question whether the TMT for example could be a predictor of speech recognition as well since the cognitive abilities involved during TMT performance are similar to those involved during speech recognition (Ellis et al., 2015). Ellis et al.:s (2015) study was built upon findings that TMT could predict speech recognition under various conditions. Adank and Janse (2010) found in a study of older (Mage = 74.1 years old)

and younger (Mage = 23.3 years old) Dutch speaking participants that TMT performance could

predict novel accent recognition, which confirmed that cognitive flexibility and executive processing are required to match distorted (distorted in terms of a unfamiliar accent) incoming speech to word representations stored in long-term memory. TMT-B scores have also been shown to correlate to RST performance and speech-in-noise recognition, in a sample of normally hearing people (Mage = 26.3 years old) (Ellis & Munro, 2013). In another study, by

Ellis and Munro (2015), with older adults (Mdnage = 75.5 years old) with hearing loss and

hearing aids both TMT-A and B performance could predict speech-in-noise performance. Ellis et al's (2015) large scale internet study also showed that both the TMT-A and B could significantly predict speech-in-noise recognition, even when the age factor was removed.

The trail making test is easily administered, and if it is proven to be a reliable tool in hearing loss assessment and treatment it can be of great use. However, as pointed out by Dugbartey, Townes and Mahurin (2000) the TMT may be of limited utility since it is believed to be sensitive to age, formal education and cultural variances. The TMT relies on the participant's

(16)

8

ability to comprehend the Latin alphabet in the TMT Part-B and illiterate and nonnative Latin alphabet users may be put at a disadvantage (Hashimoto et al., 2006, Kim, H. J, Baek, & Kim, S., 2014). To tackle these problems different versions of the TMT have been developed such as the Color Trails Test (CTT) (D'Elia, Satz, Uchiyama, & White, 1996; Maj et al., 1991; Dugbartey, Townes, & Mahurin, 2000), the Color Trails Test for children (Williams et al., 1995, Konstantopoulus, Vogazianos, Thodi & Nikopoulou-Smyrni, 2013), and Trail Making Test Black & White (TMT-B&W) (Kim et al., 2014).

In the Color Trails Test, letters have been switched out to a corresponding colored number, and the participants’ goal is to alternate between the colors and connect the numbers in ascending order, similar to the TMT. Dugbartey, Townes, & Mahurin’s (2000) study showed that in a healthy sample of 64 Turkish university students the CTT-A was of equivalence to the TMT-A which supports the results of D'Elia, Satz, Uchiyama, & White (1996) whilst CTT-2 however did not show equivalence with TMT-B, indicating that the CCT-2 might measure different cognitive skills than the TMT-B. A possible explanation according to Dugbartey et al. (2000) is the tracing of the circles differ between TMT and CTT, and the greater number of stimuli given in CTT-2 which may imply that it is a task of mere visuoperceptual ability rather than the hypothesized task-switching ability. Williams et al. (1995) suggested that the color of the encircled numbers in the Color Trails Test for children might be an interfering variable amongst children with altered neuropsychological functioning such as learning, emotional or behavioral difficulties, which is in line with Dugbartey et al.’s (2000) finding where they suggested that there can be a possible Stroop effect on CTT-2.

Kim et al. (2014) studied older adults, ranging from 50 to 80 years old, divided into three cognitively different groups (cognitively normal, with mild cognitive impairment and with Alzheimer’s disease). Their results showed that even for what they considered, in the context of elderly Korean people, highly educated (>6 years of education) cognitively normal

participants the completion rate for the TMT was 76%, and worse for people with mild

cognitive impairment (60%) and Alzheimer’s disease (37%). For participants with <6 years of education the completion rate for the TMT was 0% for all three groups. Kim et al. (2014) developed the TMT-B&W under the assumption that the completion rate would go up, which it did, but the question remains whether the TMT-B&W is still maintaining the same

psychometric properties of the TMT. The TMT-B&W-B consists of 49 encircled numbers just as in the CTT, indicating that the TMT-B&W might be a purer measure of visual perception

(17)

9

rather than task-switching. The results showed that the TMT-B&W correlated to other

neuropsychological tests related to frontal executive functions indicating that the TMT-B&W actually is in some aspect equal to the TMT. In an attempt to further explore the usage of TMT-B&W, another study was conducted (Kim, K, Jang, Baek, Kim, S, 2016). Three

versions of the TMT were used: TMT, TMT-B&W, and a Korean version of the TMT (TMT-K). The results showed yet again a low completion rate of the TMT and also a low completion rate of the TMT-K, indicating that simply switching the Latin alphabet to the Korean alphabet was not sufficient enough for use in a clinical setting. In their study the group of highly

educated Korean elderly (Meducation = 15.32 years) who knew both the Korean and English

alphabet showed difficulty of recalling the order of both the alphabet, suggesting that knowing the alphabet and knowing the order of alphabet is two separate things. Kim, K, Jang, Baek, Kim, S (2016) thus suggests that replacing the English alphabet with the Korean alphabet does not help in the assessment of detecting early stages of cognitive impairment. This speaks in favor of a non-colored and non-semantic version of the TMT in studies when controlling for language is a must.

2.5 Research questions

Earlier studies have shown that alternative versions of the TMT have been of psychometric equivalence of the TMT, however, no studies have examined the relationship between alternative versions of the TMT and speech-in-noise recognition and Reading Span Task performance. By combining strengths and weaknesses of previously created alternative TMT versions, a new version of TMT was created. This study will be unique in its sense that it uses a never previously tested version of the TMT. Very few previously conducted studies

examining the relationship between working memory, executive functions and speech-in-noise recognition have been conducted in a Swedish setting, and therefor this study may serve as a gap closer in that regard, since it not only can find new recommendations for future Swedish speech-in-noise studies, but can also possibly validate previous findings.

If a non-semantically dependent version of the TMT correlates as well as the original TMT with speech-in-noise performance and Reading Span Task performance, we would consider them of psychometric equivalence in that regard. This can indicate that the relationship

(18)

10

between TMT, RST and speech-in-noise performance is due to shared cognitive functions which are not semantically dependent. Thus the research questions for this thesis is:

1. Is the alternative version of the Trail Making Test of psychometric equivalence of the original Trail Making Test?

2. Is the cognitive abilities thought to be indexed by the Trail Making Test and Trail Making Test (alternative version) the same or similar to the cognitive functions thought to be indexed by the Reading Span Task and Speech-in-noise recognition?

(19)

11

3. Method

In this method section the design of the study, the participants and ethical aspects, materials and cognitive tests used and the procedure will be explained.

3.1 Design

The study used a between group design to control for possible training effects on the alternate forms of the Trail Making Test. 11 of the participants were given the three tests: Trail Making Test (Part A and B), Reading Span Task and Speech-in-Noise Test. 10 of the participants were given the Trail Making Test – Black & White (Part A and B), Reading Span Task, Speech-in-Noise Test. The order for the three tests were randomized for each participant. All tests were completed in a quiet room in one session lasting approximately 35 minutes.

3.2 Participants and ethics

A convenience sample of 21 university students (13 men and 8 women) was recruited by e-mail and word of mouth at Linköping University, Sweden. Participants were all informed before the test session that the study was within the area of cognitive hearing science, consisted of three different cognitive tests and it would approximately take 35 minutes to complete. All participants were native Swedish speakers with a minimum of one-year of full time study at undergraduate level, and reported that they had no documented hearing deficits. Ages varied between 21-40 years (M = 25.81, SD = 5.60). Before each test session began each participant was given an information sheet about the broader context of the study and that it comprised part of a bachelor thesis in cognitive science. After reading through the information sheet, a consent form was signed by the participant, in which they were asked to confirm that they had received enough information about the study and understood it, and that they, at any moment could cancel their participation without giving any reason.

(20)

12

3.3 Materials

Three tests were administered. The trail making test (either the original version or a newly created black and white version), a Swedish version of the Reading Span Task and a Swedish Speech-in-noise test.

3.3.1 Trail Making Test

The original Trail Making Test (Part A and B) was conducted according to standard procedure (Reitan, 1958) with paper and pen. In Part A the participants is required to connect the

encircled numbers in descending order from 1 to 25. In Part B the participant is required to connect the encircled numbers in descending numeric order as well as connecting the letters in alphabetic order, while altering between the numbers and letters, as fast as possible. The outcome measure is time taken to complete the task. If a participant were to incorrectly connect two circles in the correct order it was pointed out by the test leader, thus no errors were counted and the error rate was instead reflected in the total completion time. Before each part (A and B) were carried out, an example trial was given for the participant to try out in order to confirm that the participant had understood the task.

3.3.2 Trail Making Test – Black & White

The Trail Making Test - Black & White was a remade version of the original Trail Making Test. Whilst Part A was kept the same, in Part B the letters (eg. A-B-C) were switched out for a corresponding number (eg. A=1, B=2, C=3).Each circle were then either colored in black with the number printed in white, or colored in white with the number printed in black. In order to combat the problems with the greater number of stimuli as in the Color Trails Test (Dugbartey, Townes, & Mahurin’s, 2000) and the Korean TMT-Black & White (Kim et al., 2014) the spatial layout and number of circles were kept exactly the same but the value and color of the circles differed from the original Trail Making Test. The outcome measure is time taken to complete the task. As with the original Trail Making Test, if a participant were to incorrectly connect two circles in the correct order it was simply pointed out by the test leader, thus no errors were counted and the error rate was instead reflected in the total completion time. Before each part (A and B) were carried out an example trial was given for the participant to try out in order to confirm that the participant had understood the task.

(21)

13

3.3.3 Reading Span Task

The reading span test administered in this study was a shorter version of the Swedish RST used in Rönnberg et al., 1989. The short RST was used by Classon (2013) where the results showed that the shorter RST was comparable to the longer RST, only that it differed in scores in older adults, which possibly indicates that the longer version is more mentally demanding for older adults specifically.

A total of six blocks of three to five sentences each was presented for the participant at a 13 inch laptop screen. Every sentence consisted of three words and was presented word-by-word at a rate of 800msec per word. After each sentence the participant were to answer whether the sentence was normal or not (yes/no), the time limit to answer this question was five seconds before the first word of the next sentence appeared. At the end of each block a dialog box popped up on the screen where the participant was asked to recall either the first or last word of each sentence from that block. Participants were not pre-informed about whether they were to remember the first or last word. No time limit for this was given. The answers were typed in manually in any order, if the participant did not remember a word they were asked to simply leave it blank. A total of 24 sentences were presented in sets of 3-5 sentences each, with a total of two trials per set. In contrast to the original Reading Span Task (Daneman and Carpenter, 1980) the scoring system used in this study was the one suggested by Rönnberg et al. (1989), Lunner (2003) and Rönnberg (2003), where the score is the total numbers of words correctly recalled, in any order. It has been suggested to be a more accurate way of scoring the reading span task.

3.3.4 Speech-In-Noise test

Six blocks with ten sentences each were presented for the participant. The sentences were taken from the Swedish HINT corpus (Hällgren et al., 2006) and were the same sentences used in Ellis & Rönnberg (2014). The participant listened to the sentences using a pair of Steelseries Siberia V2 headphones. All sentences were presented at the same level for each participant: approximately 60 dB SPL. Within each block, the background noise (which was two-talker babble) varied in 3 dB steps from +12 to -15 dB in an ascending order, where the signal-to-noise ratio decreased as the participant went further into the block, making it harder to perceive the sentences at the end of the block. The sentences were presented one at a time, and after each sentence the participants were to repeat the sentence out loud to the test leader.

(22)

14

Participants were encouraged to repeat as much as they heard even in cases where they could not hear the whole sentence. Before the trial began the participant was presented with a test block to ensure that they had understood the task. The total number of correctly repeated keywords were counted as the outcome measure.

3.4 Analysis

To see test whether the data was normally distributed and that the assumptions of normality was met Shapiro-Wilks tests were conducted and the histograms and P-P plots for each variable were analyzed.

Although only one of the variables turned out to be significant on the Shapiro-Wilks tests, most other variables were close to being significant and due to the small sample size and the highly selective sample non-parametric tests were chosen.

In order to understand the relationship between the two versions of the Trail Making Test, the Reading Span Task and Speech-in-noise recognition test, non-parametric Spearman’s rho was used to examine the correlation between the scores on each test. A Mann-Whitney U Test were used to examine whether the completion time between the two versions of the Trail Making Test differed in order to interpret whether they are of qualitative difference or/and psychometric equivalence.

(23)

15

4. Results

In this section the results from the TMT-A and B, TMT-Black & White- A and B, Reading Span Task and Speech-in-noise test will be presented. Along with the results from the correlation tests.

Table 1. Descriptive statistics of the variables.

Variable N M SD Range RST 20 12.65 2.98 8-18 SIN 21 156.33 15.39 124-179 TMT-A (seconds) 11 28.91 8.829 18-47 TMT-B (seconds) 11 88.73 46.06 36-166 TMT-B&W-A (seconds) 10 23.30 7.33 17-41 TMT-B&W-B (seconds) 10 43.80 19.28 24-83

Comparison between TMT and TMT-Bl&W

The Mann-Whitney U Test for the Part A of the two TMT versions indicated that there was no significant difference between the two groups (Mdn = 24), U = 30, p = .078, r = -.38. For Part B however a significant difference between the two version were found (Mdn = 57), U = 19, p = .011, r = -.55, indicating that it took significantly longer time to complete the TMT-B compared to the TMT-Bl&W-B.

(24)

16

Correlations of TMT and TMT-Bl&W with RST and SIN

Table 2. Correlations between TMT, TMT-Bl&W, RST and SIN.

RST SIN TMT-A (Sig.) -.127 (.727) -.162 (.634) TMT-B (Sig.) -.512 (.130) -.574 (.065) TMT-Bl&W-A (Sig.) -.327 (.356) -.031 (.933) TMT-Bl&W-B (Sig.) -.495 (.145) -.575 (.082) TMT (B-A) (Sig.) -.562 (.091) -.605* (.049) TMT-Bl&W (B-A) (Sig.) -.346 (.328) -.596 (.069)

The Spearman’s correlation tests showed that no standard measure (eg. time) for participants conducting the TMT (A and B) or participants conducting TMT-Bl&W (A and B) was significantly correlated to either Reading Span or Speech-in-noise performance. The B-A difference score for the original Trail Making Test was significantly correlated to Speech-in-noise performance whilst the B-A difference for the TMT-Black & White version showed a similar pattern but no significant correlation to neither Reading Span nor Speech-in-noise performance.

Spearman’s rho was calculated between scores on the RST and SIN resulting in r(20) = .215, p = .364, showing that there was no significant correlation between the two variables.

(25)

17

5. Discussion

This discussion section is divided into two parts: one for results discussion and one for method discussion.

5.1 Result discussion

The purpose of this study was to test whether a non-semantically dependent version of the Trail Making Test was psychometrically equivalent to the original Trail Making Test, in order to try to create a culturally unbiased version of the Trail Making Test. The purpose was also to examine the relationship between Trail Making performance and working memory capacity and how it influences the ability to perceive speech in a noisy context. Two groups of

participants was assed the original TMT or TMT-Black & White, a Reading Span Test and a Speech-in-noise test. Completion time between the two TMT versions was compared to each other, and correlation between completion time, RST and SIN performance was calculated in order to examine the relationship between the three variables.

By first looking at the descriptive data (in Table 1) and trying to compare the results with previously conducted studies it is possible to see whether this study gave similar results. Tombaugh (2004) gathered a large set of sample data of TMT scores stratified by age (ranging from 18-89 years old) and education in a sample of 911 healthy Canadian individuals. To see whether the results in this study could be compared to Tombaugh’s study only those samples of similar age was selected. Since Tombaugh’s (2004) data was stratified in different age intervals resulting in different ranges and means for each group, the highest mean of all groups was selected and the highest value of minimum time and maximum time in all groups was selected to give a fair comparison. Giving the age span of 18-44 years old with TMT-A; M = 28.54, range = 12-50 and TMT-B; M = 58.46, range = 29-95. By a quick look at Table 1. we see that the normative data in Tombaughs (2004) sample is fairly comparable to both TMT-A and B and TMT-Bl&W-A and B in terms of mean and range, with the exception of the very high upper limit of TMT-B score in this study. Although the tests of normality showed now significance which cancels out possible extreme outliers, these results might be the result of some unforeseen factor such as undetected cognitive decline or unclear instructions. The significant difference in completion time between TMT-A and B, and TMT-Bl&W might possible be due to these outliers, given the small sample of this study. The B part of the TMT-Bl&W took longer to complete than corresponding A part, indicating that the condition in part

(26)

18

B is in some aspect more difficult than part A, but regarding the high upper limit of the TMT-B it is problematic to compare the completion times between TMT-TMT-Bl&W and TMT. The equivalence of the two tests is instead to be decided by comparing the versions by looking at its relations to the RST and SIN scores.

The results shows that there was no significant difference in completion time between the both groups on TMT-A and TMT-Bl&W-A, which is of no surprise considering the fact that it is the same test. On Part B however, there was a significant difference in completion time, showing that the completion time was significantly longer for the participants who conducted the original TMT version compared to those how conducted the Bl&W. Neither A nor TMT-Bl&W-A was correlated to RST or SIN scores, showing very weak correlation values and low significance levels, which in the case of TMT-A is contradictory to the findings of Ellis et al. (2015) where they showed that both TMT-A and B significantly correlated to speech-in-noise recognition. However the speech in noise test used in Ellis et al’s (2015) study was a set of di-sybyllic words in a speech-shaped background noise. Although the B parts of both TMT version did not significantly correlate to SIN, their r-values was considering the small sample very high, with small p-values. This suggests that TMT-B and TMT-Bl&W-B scores both indicate a relationship to speech-in-noise performance. This would be in line with the findings of Ellis & Munro (2013, 2015) where TMT-B performance was a significant predictor of speech-in-noise recognition. Although the SIN tests used in their studies was different, it contained sets of sentences read out by a female speaker in a background of multitalker babble, the stimuli resembled the HINT sentences (Hällgren, Larsby, & Arlinger, 2006) since both stimuli utilized whole sentences and the background noise was actual people speaking and not speech-shaped noise as in Ellis et al. (2015).

Sánchez-Cubillo et al. (2009) suggested the B-A difference as a pure measures of executive function and was thus utilized in this study. Ellis and Munro (2013) suggested, along with previous findings by Engle (2002), that a greater working memory capacity was the result of better attentional control rather than storage capacity, hence the interpretation that the Reading Span Task would be a measure of executive function rather than pure working memory. The TMT B-A score of this study was significantly correlated to SIN performance, and was strongly correlated (close to significant) to RST. Taken all the points above together this would strengthen both the argument that the RST is a measure of executive function and that the A difference score is a plausible measure of executive function. For TMT-Bl&W the

(27)

B-19

A measure did not significantly correlate to RST or SIN performance. Although the value seemed to follow the same pattern as for the original TMT version, the correlation was not as strong (p = .069). This might indicate that the TMT-Bl&W does not measure executive

function or working memory capacity to the same extent the original TMT does, or that it does not measure the same executive functions.

Probably the most shocking results it the weak correlation between Reading Span and speech-in-noise scores. This contradicts most studies reviewed in this thesis (Akeroyd, 2008;

Rönnberg et al., 1989; Rönnberg et al., 2013). A possible explanation to this can be the lack of statistical power, since the correlation was not very weak and not totally insignificant.

Füllgrabe & Rosen (2016) however suggested that in younger populations with normal hearing thresholds working memory capacity is not a strong predictor of speech-in-noise recognition, at least not when measured by the Reading Span Task.

5.2 Method discussion

One obvious weakness of this study is the small sample size. Many of the correlation scores were moderate to high, indicating a possible effect, but due to low statistical power did not show significance. The sample was also a convenience sample of undergraduate students which makes you question the external validity. The sample may not be a correct

representation of population as a whole since both age and education level have been shown to effect TMT performance (H. J. Kim, Baek, & Kim, 2014; K. Kim & Jang, 2016;

Tombaugh, 2004). As pointed out in the section above, a few outliers may have affected the result, but due to the small sample none of the scores was removed. Another aspect to mention is the fact that the participants had no self-reported hearing deficit, which does not exclude the possibility that they may have had some kind of hearing deficit. Participants was not asked whether they had any self-reported or clinically diagnosed cognitive impairment, which makes it unpredictable to say to what degree the group was homogenous in terms of cognitive abilities.

Experimentally, many of the studies that have been comparing different versions of the TMT have used a within-group design ( Williams et al., 1995 ,Dugbartey et al., 2000; H. J. Kim, Baek, & Kim, 2014; K. Kim & Jang, 2016). This may cause either a possible training or transfer effect as suggested by Dugbartey et al. (2009). In order to combat these two threats to validity a between-groups design was chosen. This of course requires more participants but is

(28)

20

weighed up by the elimination of named threats of previous studies. A possible way to validate the results of this study could be to conduct a repeated measures or a longitudinal within-group experiment to see whether participants performed different under similar conditions. Since the Black and White version has never been tested previously a study with larger sample and a test-retest design would be suitable.

The short version of the Reading Span Task (Rönnberg et al., 1989) was used since Classon (2013) found that it was more or less equivalent to the longer version. The assumption about the equivalence is based on this finding only which does not cancel out the possibility that the use of the longer version would yield different results. The use of a Swedish RST can make it more difficult to compare the results to other international studies. However, Rönnberg et al.:s (1989) version have been used extensively and been translated to other languages as well (Rudner, Foo, Sundewall-Thorén, Lunner, & Rönnberg, 2008).

Something that speaks on favor of the ecological and external validity of this study is the use of sentences from the Swedish HINT corpus (Hällgren et al., 2006). The Swedish HINT corpus was created with the intention of being a clinically usable hearing test tool comparable to the widely used American English HINT sentences (Nilsson, Soli, & Sullivan, 1994) and as a complementary tool to the Swedish Hagerman sentences (Hagerman, 1982). This might make the results more comparable to other national as well as international studies in the field.

Due to the fact that this was the first time the TMT-Bl&W was tested it cannot be compared to similar longitudinal studies, which of course makes you question its validity. In this study only two other cognitive tested were used to compare the two TMT versions, thus it may not be satisfactory enough to say that the TMT-Bl&W is not of psychometric equivalence with TMT since this is only based on the comparison of two other cognitive tests beyond the tests themselves. Further suggestion would be to include more cognitive tests for referential data.

(29)

21

6. Conclusion

The results points at being both in line with previous studies as well as showing some contradictory findings. The two TMT versions significantly differed in completion time and the original TMT version was more correlated to RST and SIN performance compared to the TMT-Bl&W. This indicates that the cognitive functions indexed by the original TMT differed from the functions indexed by the TMT-Bl&W. This could however be an effect of many different variables, such as lack of statistical power, a previously untested version of the Trail Making Test, lack of normative data for alternative Trail Making versions, and too few referential cognitive tests, thus one cannot say for sure whether or not the TMT and TMT-Bl&W is of psychometric equivalence. The B-A difference score is suggested to be the best measure of executive function and the greatest predictor for Reading Span and speech-in-noise performance. Suggestion for further studies is to test a larger sample for the TMT-Black & White, additional cognitive tests and using a repeated measures design.

(30)

22

7. References

Adank, P. & Janse, E. (2010) Comprehension of a novel accent by young and elderly listeners. Psychology and Aging, 25, 736-740.

Akeroyd, M. A. (2008). Are individual differences in speech reception related to individual differences in cognitive ability? A survey of twenty experimental studies with normal and hearing- impaired adults. International Journal of Audiology ISSNOnline) Journal

International Journal of Audiology, 47, 1499–2027.

Anderson, MC, & Levy, B. (2009). Suppressing unwanted memories. Current Directio ns in Psychological Science, 18, 189–94.

Arbuthnott, K., & Frank, J. (2000). Trail Making Test, Part B as a Measure of Executive Control: Validation Using a Set-Switching Paradigm. Journal of Clinical and

Experimental Neuropsychology, 22(4), 518–528.

Arlinger, S., Lunner, T., Lyxell, B., Pichora-Fuller, M. K., Pichora, M. K., & Pichora-Fuller, M. K. (2009). Linköping University Post Print The emergence of cognitive hearing science. The Emergence of Cognitive Hearing Science. Scandinavian Journal of

Psychology, 5(50), 371–384.

Atkinson, R.C., & Shiffrin, R.M. (1968). Human memory: A proposed system and its control processes. In K. W. Spence & J. T. Spence (Eds.), The psychology of learning and motivation: Vol 2. Advances in research and theory. New York: Academic Press. Baddeley, A. (1996). Exploring the central executive. Q. J. Exp. Psychol. A, 49, 5–28. Baddeley, A. (2000). The episodic buffer: a new component of working memory?. Trends

in cognitive sciences, 4(11), 417-423.

Baddeley, A. (2012). Working Memory: Theories, Models and Controversies. Annual review of Psychology 63, 1-23.

Baddeley, A., & Hitch, G. (1974). Working memory. In G. . Bower (Ed.), The psychology of

learning and motivation: Advances in research and theory (pp. 47–89). New York:

Academic Press.

Bower, G. H. (2000). A brief history of Memory Reasearch. In Craik, F., Tulving, E. (Red.), The Oxford handbook of memory. New York: Oxford University Press.

Broadbent, D. (1958). Perception and Communication. London: Pergamon Press

(31)

23

adaptability to environmental changes in dynamic complex problem-solving tasks.

Ergonomics, 46(5), 482–501.

Classon, E. (2013). Representing sounds and spellings Phonological decline and

compensatory working memory in acquired hearing impairment. Linköping University.

Conway, A. Kane, M., Bunting, M., Hambrick, Z., Wilhelm, O., Engle, R. (2005). Working memory span tasks: A methodological review and user's guide. Psychonomic Bulletin & Review, 12(5), 769-786.

Craik, F., Lockhart, R. (1972). Levels of Processing: A Framework for Memory Research. Journal of Verbal Learning and Verbal Behavior, 11, 671-684.

Daneman, M., & Carpenter, P. A. (1980). Individual Differences in Working Memory and Reading. Journal Of Verbal Learning and Verbal Behavior, 19, 450–466.

Daneman, M., & Merikle, P.M. (1996). Working memory and language comprehension: A meta-analysis. Psychonomic Bulletin & Review 3(4) 422-433.

Diamond, A. (2013). Executive Functions. Annu. Rev. Psychol, 64, 135–68.

Dugbartey, A. T., Townes, B. D., & Mahurin, R. K. (2000). Equivalence of the Color Trails Test and Trail Making Test in Nonnative English-Speakers. Archives of Clinical

Neuropsychology, 15(5), 425–431.

D'Elia, F., Satz, P.,Uchiyama, C.L.,White, T. (1996). Color Trails Test. Professional manual, Psychological Assessment Resources, Odessa, FL.

Ellis, R., Molander, P., Rönnberg, J., Lyxell, B., Andersson, G., & Lunner, T. (2015). Predicting Speech-in-Noise Recognition from Performance on the Trail Making Test: Results from a Large-Scale Internet Study. Ear and Hearing, 37(1), 73-79.

Ellis, R., & Munro, K. (2013). Does cognitive function predict frequency compressed speech recognition in listeners with normal hearing and normal cognition? International Journal

of Audiology, 52(1), 14–22.

Ellis, R., & Munro, K. (2015). Predictors of aided speech recognition, with and without frequency compression, in older adults. of aided speech recognition, with and without frequency compression, in older adults. International Journal of Audiology, 7(54), 467– 75.

Ellis, R., & Rönnberg, J. (2014). Cognition and Speech-In-Noise Recognition: The Role of Proactive Interference. Journal of the American Academy of Audiology, 25(10), 975–982. Ellis, R., & Rönnberg, J. (2015). How does susceptibility to proactive interference relate to

speech reocgnition in aided and unaided conditions? Frontiers in Pschology, 6, 1017. Engle, R. W. (2002). Working Memory Capacity as Executive Attention. Current Directions

(32)

24

in Psychological Science, 11(1), 19–23.

Füllgrabe, C., & Rosen, S. (2016). On The (Un)importance of Working Memory in Speech-in-Noise Processing for Listeners with Normal Hearing Thresholds. Frontiers in

Psychology, 7, 1268.

Gomez, P., Ratcliff, R., & Perea, M. (2007). A model of the go/no-go task. Journal of

Experimental Psychology. General, 136(3), 389–413.

Hashimoto, R., Meguro, K., L., Kasai, M., Ishii, H., Yamaguchi, S. (2006). Effect of age and education on the Trail Making Test and determination of normative data for Japanese elderly people: The Tajiri Project. Psychiatry and Clinical Neurosciences, 60(4), 422-428.

Hagerman, B. (1982). Sentences for testing speech intelligibility in noise. Scandinavian

Audiology, 11(2), 79–87.

Hofmann, W., Friese, M., & Strack, F. (2009). Impulse and self-control from a dual-systems perspective. Perspectives on Psychological Science, 4, 162–76.

Hällgren, M., Larsby, B., & Arlinger, S. (2006). A Swedish version of the Hearing In Noise Test ( HINT ) for measurement of speech recognition. International Journal of

Audiology, 45, 227–237.

Kane, M. J., Bleckley, M. K., Conway, A. R. A., Kane, R. W. E., & Bleckley, M. J. (2001). A Controlled-Attention View of Working-Memory Capacity. Journal of Experimental

Psychology: General, 130, 169–183.

Kiessling, J., Pichora-Fuller, M. K., Gatehouse, S., Stephens, D., Arlinger, S., Chisolm, T., … von Wedel, H. (2003). Candidature for and delivery of audiological services: special needs of older people. International Journal of Audiology, 42(sup2), 92–101. Kim, H. J., Baek, M. J., & Kim, S. (2014). Alternative Type of the Trail Making Test in

Nonnative English-Speakers: The Trail Making Test-Black & White. PLoS ONE, 9(2). Kim, K., & Jang, J. W. (2016). A Comparison of Three types of Trail Making Test in the

Korean Elderly: Higher Completion Rate of Trail Making Test-Black and White for Mild Cognitive Impairment. Journal of Alzheimer’s Disease & Parkinsonism, 6(3).

Lehto, J. E., Juujärvi, P., Kooistra, L., & Pulkkinen, L. (2003). Dimensions of executive functioning: Evidence from children. British Journal of Developmental Psychology,

21(1), 59–80.

Lin, F. R., & Albert, M. (2014). Hearing loss and dementia - who is listening? Aging &

Mental Health, 18(6), 671–3.

(33)

25

B. (2011). Hearing Loss and Incident Dementia. Archives of Neurology, 68(2), 260–266. Lyxell, B., & Rönnberg, J. (1989). Information-processing skill and speech-reading. British

Journal of Audiology, 23(4), 339–47.

Lyxell, B., & Rönnberg, J. (1991). Word discrimination and chronological age related to sentence-based speech-reading skill. British Journal of Audiology, 25(1), 3–10.

Lyxell, B., & Rönnberg, J. (1992). The relationship between verbal ability and sentence-based speechreading. Scandinavian Audiology, 21(2), 67–72.

Lyxell, B., Rönnberg, J., & Samuelsson, S. (1994). Internal speech functioning and

speechreading in deafened and normal hearing adults. Scandinavian Audiology, 23(3), 179–85.

Maj, M., Satz, P., Janssen, R., Zaudig, M., Starace, F., D’Elia, L., … Ndetei, D. (1994). WHO Neuropsychiatric AIDS study, cross-sectional phase II. Neuropsychological and

neurological findings. Archives of General Psychiatry, 51(1), 51–61.

Milner, B. (1963). Effects of Different Brain Lesions on Card Sorting. Archives of Neurology,

9(1), 90.

Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The Unity and Diversity of Executive Functions and Their Contributions to Complex Frontal Lobe’’ Tasks: A Latent Variable Analysis. Cognitive Psychology, 41, 49–100.

Nilsson, M., Soli, S. D., & Sullivan, J. A. (1994). Development of the Hearing In Noise Test for the measurement of speech reception thresholds in quiet and in noise. The Journal of

the Acoustical Society of America, 95(2), 1085–1099.

Pichora-Fuller, M. K. (2007). Audition and cognition: What audiologists need to know about listening. In C.Palmer & R.Seewald (eds.) Hearing care for adults (pp. 71–85). Stäfa, Switzerland: Phonak.

Posner MI, DiGirolamo GJ. 1998. Executive attention: conflict, target detection, and cognitive control. In The Attentive Brain, ed. R Parasuraman, pp. 401–23. Cambridge, MA: MIT Press.

Reitan, R. M. (1958). Validity of the Trail Making Test as an indicator of organic brain damage. Perceptual and Motor Skills, 8(7), 271.

Rudner, M., Foo, C., Sundewall-Thorén, E., Lunner, T., & Rönnberg, J. (2008). Phonological mismatch and explicit cognitive processing in a sample of 102 hearing-aid users.

International Journal of Audiology ISSNOnline) Journal, 47(2), 91–98.

(34)

26

and “handicap age.” European Journal of Cognitive Psychology, 2(3), 253–273. Rönnberg, J. (2003). Cognition in the hearing impaired and deaf as a bridge between signal

and dialogue: a framework and a model. International Journal of Audiology. 42, 68–76. Rönnberg, J., Arlinger, S., Lyxell, B., & Kinnefors, C. (1989). Visual evoked potentials:

relation to adult speechreading and cognitive function. Journal of Speech and Hearing

Research, 32(4), 725–35.

Rönnberg, J., Lunner, T., Zekveld, A., Sörqvist, P., Danielsson, H., Lyxell, B., … Rudner, M. (2013). The Ease of Language Understanding (ELU) model: theoretical, empirical, and clinical advances. Frontiers in Systems Neuroscience, 7, 31.

Rönnberg J., Rudner M., Foo C., Lunner T. (2008). Cognition counts: a working memory system for ease of language understanding (ELU). Int. J. Audiol., 47(2), 99–105.

Rönnberg J., Rudner M., Lunner T., Zekveld A. A. (2010). When cognition kicks in: Working memory and speech understanding in noise.

Sánchez-Cubillo, I., Periáñez, J., Adrover-Roig, D., Rodríguez-Sánchez, J., Ríos-lago, M., Tirapu, J., & Barceló, F. (2009). Construct validity of the Trail Making Test: Role of task-switching, working memory, inhibition/interference control, and visuomotor abilities. Journal of the International Neuropsychological Society, 15, 438–450. Stenbäck, V., Hällgren, M., & Larsby, B. (2016). Executive functions and working memory

capacity in speech communication under adverse conditions. Speech, Language and

Hearing, 19(4), 218–226.

Sternberg, R., Sternberg, K. (2012). Cognition (6 th ed.). Canada: Cengage Learning. Sutherland, N. S. (1968). Outlines of a theory of visual pattern recognition in animals and

man. Proceedings of the Royal Society of London. Series B, Biological Sciences,

171(1024), 297–317.

Theeuwes, J. (1991). Exogenous and endogenous control of attention: the effect of visual onsets and offsets. Perception & Psychophysics, 49(1), 83–90.

Tombaugh, T. N. (2004). Trail Making Test A and B: Normative data stratified by age and education. Archives of Clinical Neuropsychology, 19, 203–214.

Turner, M. L., & Engle, R. W. (1989). Is Working Memory Capacity Task Dependant? Journal of Memory and Language 28. 127-154.

Uhlmann, R. F., Larson, E. B., Rees, T. S., Koepsell, T. D., & Duckert, L. G. (1989). Relationship of Hearing Impairment to Dementia and Cognitive Dysfunction in Older Adults. JAMA: The Journal of the American Medical Association, 261(13), 1916. van Rooij J.C.G.M. & Plomp R. (1990). Auditive and cognitive factors in speech perception

(35)

27

by elderly listeners. II. Multivariate analyses. J Acoust Soc Am, 88, 2611-2624. Williams, J., Rickert, V., Hogan, J., Light, R. (1995). Children's color trails. Archieves of

Executive function, working memory and speech-in-noise recognition – Comparing a non-semantic black and white version of the Trail Making Test to the original Trail Making Test