The impact of induced anxiety on updating in working memory

Размещено на http://www.allbest.ru/

The Government of the Russian federation

Federal State Autonomous Educational Institution of Higher Professional Education

National Research University - Higher School of Economics

Faculty of Social Sciences, School of Psychology,

Master's program

“Cognitive sciences and technologies: from neuron to cognition”

Master thesis

«The Impact of Induced Anxiety on Updating in Working Memory»

Lyusin Dmitry Vladimirovich



Table of Contents

Introduction

1. Aim

1.1 Hypothesis

2. Literature Review

2.1 Definition and Classification of Anxiety

2.2 State and Trait Anxiety

2.3 Attentional Control Theory

2.4 Moderators of Attentional Bias

2.5 Working Memory

3. Relationship Between Working Memory and Anxiety

3.1 Anxiety and Working Memory Performance

3.2 Stress and Working Memory

3.3 Anxiety and Working Memory Updating

3.4 Trait Anxiety and Working Memory

4. Limitations of Prior Research

5. Methods

5.1 Participants

5.2 Materials

5.3 Tasks

5.4 Procedure

6. Data Analysis and Results

6.1 STAI-40 and STAI-6

7. Discussion and Summary

Conclusion

References



Introduction

Cognitive psychology has progressed steadily in multiple domains of cognitive processes over the last few decades. Despite the progress, few domains have not been explored extensively leading to biases in theoretical understanding of certain phenomena. One of the domains less explored is the relationship between state anxiety and updating. Cognitive deficit is a well-recognized component of anxiety. Particularly, anxiety restricts several processes of the working memory such as updating, attentional inhibition, and attentional shift. Most research studies that have investigated the role of anxiety in working memory measures have dealt with trait anxiety using tasks that are storage oriented. There is a lack of research on the updating especially under the influence of state anxiety. The aim of the present study is to explore the relationship between induced anxiety and working memory updating. To induce anxiety, a variety of mood manipulation techniques will be used such as autobiographical recall, videos, and music. State and trait anxiety will be measured with State-Trait Anxiety Inventory (STAI) (Speilberger, Gorsuch, & Lushene, 1970). The tasks used in this study are 2-back non-affective task which is the regular version with alphabets as the stimuli and a modified 2-back affective task where the stimuli are faces displaying a range of emotions. The results of the study would help reach a distinct understanding of how state anxiety impairs updating.



1. Aim

To find out how state anxiety impacts working memory updating. Specifically, it would require finding the effect of induced anxiety on updating.

1.1 Hypothesis

Participants with high levels of anxiety will be less effective in the updating of affective and non-affective information with longer reaction time and reduced accuracy compared to participants with low levels of anxiety. If the hypothesis is true, it would suggest that stressors of daily life that cause anxiety affect attentional and working memory resources that lead to impairment of updating.



2. Literature Review

2.1 Definition and Classification of Anxiety

Anxiety is a state associated with heightened vigilance and greater sensory sensitivity when a conflict or uncertainty is encountered (Grillion, 2002; Grupe & Nitschke, 2013; Eysenck et al., 2007). Sigmund Freud defined anxiety as a fundamental phenomenon which is the primary problem of neurosis (Freud, 1936). It is characterized with an increased emphasis on negative stimuli, attentional bias, and limited or no control over worrisome thoughts (Matthews & Wells, 1996). The psychological characteristics of anxiety also have physiological manifestations such as rapid heart rate, increased blood pressure, muscular disturbances like tremors, and physical weakness (Kielholz, 1990). The interpretation of anxiety can differ across cultures but the general interpretation is that anxiety is dangerous in situations where there is psychological or physical danger while anxiety is deemed normal in certain non-threatening situations. However, when anxiety occurs in the absence of potential physiological or psychological danger or the intensity of emotional reaction is inconsistent with the danger, it is indicative of neurotic or clinical anxiety (Kielholz, 1990). The concept of anxiety evolved in the twentieth century with significant contributions from Freud. He proposed the danger signal theory which posits that danger can come from two potential sources: one's internal impulses, and the external world. He further applied this theory to anxiety proposing that when the source of danger is the external world, objective anxiety is experienced. Objective anxiety is synonymous with fear which is an emotional state aroused by harm from the danger source.

2.2 State and Trait Anxiety

The classification of state anxiety evolved from the concept of objective anxiety. Krause (1961) stated that anxiety can be inferred from physiological signs, molar behavior, introspective verbal reports, response to stress, and clinical intuition. Based on these response patterns, Marin (1961) differentiated state anxiety suggesting that it can be defined by physiological signs, behavioral signs, and introspective verbal reports. Cattell and Scheier applied multivariate techniques to measure anxiety and classified two independent factors of anxiety which were state and trait. Schacter (1964) further added that socially determined cognitions are also involved in triggering stimuli which lead to a cognitive-physiological experience. State anxiety is a multifaceted emotional state triggered by an internal or external threat. It can be defined as an acute condition that is time specific (Caselli, Reiman et al., 2004). It is characterized by cognitive components such as worry and attentional bias, emotional components such as nervousness and negative feelings, physiological components such as arousal, and behavioral components such as withdrawal or avoidance (Lang, 1993). Trait anxiety is a general or chronic tendency to respond to stressful stimuli (Caselli, Reiman et al., 2004). It is characterized by a tendency to experience high levels of anxiety below the clinical level. Individuals who have high trait anxiety are usually novelty and risk avoidant ( Maner et al., 2007). Although high trait anxiety does not suggest the possibility of an anxiety disorder, it still remains a key vulnerability (Plehn & Peterson, 2002).

Psychological defenses.

When high levels of state anxiety is experienced, a coping behavior is triggered which acts as a response to the perceived threat. Individuals who have high trait anxiety are likely to experience higher levels of state anxiety compared to those with low trait anxiety. If the threat is psychological, repression is a common defence mechanism. Repression can lead to modification or distortion of thoughts, memories, and feelings which were previously triggered with a threat. Denial is another method that defends state anxiety by eliminating the perceived threat from awareness. Denial is a more primitive and drastic defense relative to repression which can also lead to the avoidance of potential danger. However, both denial and repression have been reported to reduce state anxiety by altering the perception of a threat. Nevertheless, these defense mechanisms are not completely efficient because they can become maladaptive since the cause of anxiety remains unchanged even if anxiety levels decrease.

Assessment of state and trait anxiety.

Projective techniques commonly used to evaluate anxiety are Thematic Apperception Test and Rorschach Inkblot Test. Questionnaires and self-report psychometric inventories have also been extensively used in research. One of the first scales created was the Manifest Anxiety Scale (MAS) that contains 50 questions on trait anxiety (Taylor, 1953). The Anxiety Scale Questionnaire (ASQ) was developed a decade after the MAS which was a standardized assessment of anxiety used for clinical purposes (Cattell & Scheier, 1963). MAS and ASQ were almost equivalent in terms of reliability but differed in terms of method of construction, item format, and concepts of anxiety. However, neither tests were designed with the distinction between state and trait anxiety. The Affect Adjective Check List (AACL) was created by Zuckerman and his colleagues that measures state and trait anxiety separately (Zucherman & Lubin, 1965; Zuckerman & Biase, 1962). The development of these scales later led to the most popular, standardized, and reliable inventory known as the State-Trait Anxiety Inventory (STAI) (Speilberger, Gorsuch, & Lushene, 1970). STAI is based on Freud's danger signal theory and Cattell and Scheier's state and trait anxiety conceptions (Cattell & Scheier, 1963). The questions were constructed from MAS, ASQ, and the Welsh Anxiety Scale where each question was tested for its validity, internal consistency, and psychometric properties. After testing the questionnaire on 2000 college students, 20 questions for state anxiety and 20 questions for trait anxiety were finalized. Anxiety has profound consequences that impact emotional states, social functioning, cognitive functioning, and physical health. For the sake of the present study, the impact of anxiety on cognitive functioning will be particularly emphasized on.



2.3 Attentional Control Theory

The attentional control theory proposed by Eysenck et al. (2007) posits that anxiety impairs two central executive functions associated with attentional control: shifting and inhibition. Shifting of attention is a process which is the ability to shift attention from one task to another based on the given circumstance. Inhibition, on the other hand, is the ability to regulate automatic response, dominant response, or inhibit responses. Anxiety limits the extent to which automatic responses or inhibition are executed. Similarly, in shifting, anxiety enhances the attentional shift which allows a greater attention to the stimuli or threat. This suggests that attentional bias plays an important role in how anxiety impairs cognitive functioning.

Understanding the relationship between anxiety and attentional control is imperative in learning how it influences updating. Since updating is associated with replacing irrelevant information with relevant information, attention is involved in selecting the relevant information and also in selectively eliminating the irrelevant ones. Additionally, as anxiety impacts attentional control, it is likely that attentional resources engaged in the updating process would also be impacted. To obtain a distinct understanding of attentional bias, an elaboration of its components is required. It can be explained in terms of three processes: disengaging attention, attentional avoidance, and facilitated attention. The following section discusses how attentional bias occurs under the influence of anxiety.

Disengaging attention.

Difficulty in disengaging or shifting attention from the threat leads to an impairment of the ability to switch attention between tasks or to another task away from the threat. In order to assess disengaging attention, researchers have used more than one task. Several studies have investigated how disengagement occurs by using spatial cueing tasks (Van Damme,Crombez et al., 2006; Fox, Russo, & Dutton, 2002 ). Dot probe task has also been used to show that anxious individuals have difficulty in disengaging attention from the threat (Salemnik et al., 2007; Koster et al., 2004). Visual search tasks have also been used to test difficulty in attentional disengagement. Most studies have reported consistent results that showed participants have had difficulty in disengaging attention when they were anxious (Lipp & Waters, 2007; Juth, Lundqvist et al. , 2005; Miltner, Krieschel et al., 2004 ). However, despite the consistent results of various studies, there are conflicting opinions. Pflugshaupt et al. (2005) did not find difficulty in attentional disengagement when participants were anxious using the visual search task. Additionally, delayed disengagement when a potential threat is present could occur due to a generic slowdown effect (Algom et al., 2004). Hence, Mogg et al. (2008) have put forth an argument according to which the evidence on delayed disengagement is not unambiguous as they suspect a confound between the generic slowdown that occurs on an encounter with a threat and delayed disengagement.

Attentional avoidance.

Attentional bias occurs when attentional avoidance occurs. In this process, the attention is assigned to a location opposite to that of the threat in a preferential manner, thus indicating avoidance (Koster et al., 2004). Koster et al. (2005) investigated attentional bias through dot probe task and found that there was a congruency effect in the reaction times. They found a shorter reaction time in congruent trials compared to that of incongruent trials. However, the conditions of the experiment were precisely maintained with the stimulus duration being 500 ms and the participants had high trait anxiety.

When the stimulus duration was increased to 1250 ms, the participants with high trait anxiety showed attentional avoidance when threatening cues were presented and the congruency effect was absent. This experiment has been replicated several times where similar results were obtained (Mogg, Holmes et al., 2004; Koster et al., 2006). Additionally, attention avoidance to threatening stimuli has also been measured directly in terms of eye fixations ( Garner et al., 2006). Calvo and Avero (2005) conducted an experiment where emotional and nonemotional pictures were shown to participants with high and low trait anxiety for three seconds. The results of the showed that participants with high anxiety displayed preferential attention. Additionally, participants with high anxiety had lower frequency of fixation and shorter viewing time towards pictures of harmful stimuli. Hence, individuals with high trait anxiety or under anxiety conditions tend to divert their attention from harmful stimuli by reducing their fixation frequency or reducing the viewing time.

Facilitated attention.

Facilitated attention or attentional orientation is the degree to which attention is being directed towards the stimuli. Some studies have attempted and failed to show evidence of facilitated attention under anxiety by using spatial cue tasks ( Amir, Elias, Klumpp, & Przeworski, 2003; Fox, Russo, & Dutton, 2002). Koster and colleagues (2006) induced anxiety with precise conditions by showing highly threatening pictures with a 100 ms interval between each stimulus and among individuals who already have high trait anxiety. Furthermore, only when stimuli were presented subliminally in dot probe tasks, facilitated attention was found (Carlson & Reinke, 2008). Most studies that failed to find facilitated attention did not manipulate the stimulus presentation conditions. For example, Van Damme et al. (2006) used spatial cueing tasks with classical conditioning paradigm but manipulated the stimuli by pairing aversive stimuli with neutral ones. The results of the study successfully reported facilitated attention. Therefore, although there is a mixed view whether facilitated attention is indeed a component of attentional bias, the problem seems to be lying in the experimental conditions. Based on findings from previous studies, it can be concluded that facilitated attention towards a threatening stimulus occurs when the intensity of the stimulus is high and the duration of the stimulus is less or equal to 100 milliseconds (Cisler & Koster, 2010). Thus, attentional control leads to difficulty in disengaging attention from a threatening stimuli whereas emotion regulation tends to mediate attentional avoidance from emotional or threatening stimuli.



2.4 Moderators of Attentional Bias

Attentional control.

The role of attentional control in attentional bias has been under research in the past few decades (Eysenck et al., 2007; Mathews & Wells, 2000). Attentional control is defined as the difference variable which characterizes an individual's ability to control their attentional allocation. People with high trait anxiety tend to have poor attentional control which has been identified through related attentional disengagement from threat (Derryberry & Reed, 2002). Another research study found that poor attentional control among individuals have difficulty in disengaging attention from emotional faces that have been rapidly presented with a 100 ms duration ( Peers & Lawrence, 2009).

Emotion regulation.

Emotion regulation also plays a role in regulating attention towards threat. The processes that influence when and how emotions are experienced form emotion regulation strategies (Gross, 1998). One of the mechanisms of emotion regulation has been proposed to be attentional allocation. In situations where individuals desire to reallocate their attention purposefully to a neutral stimuli in order to avoid an unpleasant or threatening stimuli is governed by emotion regulation processes (Sheppes & Meiran, 2008; Van Dillen & Koole, 2007 ).

Although both emotion regulation and attentional control have a role in moderating attentional bias, they work in different ways. People who have better attentional control have less difficulty in disengaging their attention from a threatening stimuli compared to those who have poor attentional control. However, in the case of emotional regulation, it requires more than a person's ability to regulate their emotions. For example, a person's coping mechanism in response to negative emotion can influence their emotional regulation and that can differ across individuals. Additionally, certain emotion regulation strategies like reappraisal could be more effective than other strategies (Feldner et al., 2003; Gross & Levenson, 1997). Therefore, this would lend an advantage to individuals who cope with negative emotions through reappraisal.

Research findings suggest that attentional biases occur in the form of delayed disengagement from a threatening stimuli and attentional avoidance in late processing stages whereas facilitated attention occurs in the early stages when the stimulus duration is short but the threat intensity is high.

2.5 Working Memory

Working memory is a short-term storage system with a limited capacity that has multiple components (Chai et al, 2018). The term “working memory” was used for the first time by Miller et al. (1960). Diamond and Wright (2014) describes working memory as an executive function that golds information and enables working out the information mentally. While there are various paradigms to describe working memory, one of the most prominent models is Badeley's multicomponent working memory model (Badeley, Logie, Bressi et al., 1986). This model entails three components: phonological loop, visuospatial sketchpad, central executive, and episodic buffer (Baddeley, 2000). There are various domains of working memory such as maintenance of information, updating information, inhibitory controls ( Engle & Kane, 2004), and executive processes which depend on executive control efficacy (D'Esposito & Postle, 2015). In this study, two executive functions, inhibition and shifting, and updating will be focused on.

Inhibition.

The inhibition function of the working memory enables individuals to maintain attention on the current task and prevent stimuli and responses that are irrelevant to the current task (Friedman & Miyake, 2004). One of attentional control theory's main predictions is that anxiety disrupts inhibition of the working memory, that is, it hinders the suppression of information that is irrelevant to the current task. Since worry is a component of anxiety that demands attention to threatening stimuli, attentional inhibition is predicted to be involved in cognitive tasks when an individual is undergoing stress and worry (Moran, 2016). This effect has been observed in the presence as well as the absence of threat distractors.

In antisaccade tasks, participants with higher anxiety have shown to initiate eye movement slowly when moving it away from a peripheral target in comparison to that of participants with low anxiety. In a prosaccade trial, however, the eye movement to look towards the target was not slower among individuals with high anxiety since it does not involve attention inhibition (Derakshan, Smyth, & Eysenck, 2009). Thus, anxiety has been repeatedly reported to reduce the ability to inhibit attention successfully in task-irrelevant items. Furthermore, since anxiety impacts inhibitory control on attention, an anxious individual becomes less capable of limiting irrelevant information from getting a greater access to working memory, thus, reducing the working memory capacity for current tasks.

Shifting.

Shifting enables an individual to allocate the attention to a stimulus that is relevant to the current task. Attentional control theory maintains that anxiety reduces the ability to shift attention in a flexible way among task demands. This prediction of Eysenck and colleagues (2007) has been investigated directly and indirectly in other studies that provide supporting evidence. One of the direct ways of assessing how anxiety impacts attentional shift is offered by task shifting paradigms. The performance on blocks of trials of tasks that are subject to switching gives a measure of attentional shift (Miyake et al., 2000). Monsell (2003) demonstrated that in task switching procedures, a switching cost is involved which impacts performance.

Derakshan, Smyth, & Eysenck (2009) conducted a study where task switching procedures were used to test the interaction of attentional shift and state anxiety. When the complexity of the task was increased, the performance of low anxious participants did not change significantly unlike that of high anxiety participants. The authors concluded that individuals with high state anxiety are likely to perform poorly in task-switching paradigms compared to single task conditions whereas the switching cost is minimal among individuals with low state anxiety.

The influence of anxiety on attentional shift has also been tested indirectly through Wisconsin Card Sorting Task where participants who were anxious had longer reaction times and made more errors (Caselli et al., 2004). The reduced performance in task when switching from an antisaccade task to a prosaccade trial was observed in participants who had high anxiety (Ansari et al., 2008). Ansari, Derakshan, & Richards (2008) recently conducted a study where antisaccade and prosaccade tasks were intermixed. Participants who had high anxiety showed slower response when the inter-stimulus was short in both the prosaccade and the antisaccade trials. Additionally, individuals who are prone to high anxiety tend to shift their attention to safe cues compared to neutral cues which could be attributed to their attentional process while encoding (Lundh et al., 1998).The findings of the study imply that individuals who have high anxiety use a compensatory strategy greater than that of individuals with low anxiety in order to maintain their performance. Thus, although converging evidence on the interplay of anxiety and attentional shifting are scarce, findings from prior studies show consistent results. Therefore, it can be concluded that anxiety slows down attentional shifting and increases reaction times.

Updating.

Updating is a process in which information is monitored and encoded according to the relevance of the task at hand. The information to be encoded replaces irrelevant information that is no longer required (Morris & Jones, 1990). It has been suggested that updating might involve temporal tagging to keep a track of relevant and irrelevant information (Jonides & Smith, 1997). It is important to note that updating is not the same as storing and maintaining information. In other words, updating is not limited to passively storing task-relevant information because it also entails manipulating the contents of WM (Lehto, 1996). Behavioral research has inconsistent findings and neuroscientific works are limited in this area. According to the attentional control act, it has been presumed that stressful conditions can lead to anxiety that impairs the updating system (Berggren & Derakshan, 2013). A handful of studies have investigated the impact of anxiety in cognitive tasks that involve working memory processes.

An event-related potential study by MacNamara, Ferri, and Hajcak (2003) found that anxiety was associated with increased error related negativity when participants performed Stroop tasks. This effect has also been reported by other studies (Aarts & Pourtois, 2010; Olvet & Hajcak, 2009). In tasks that require selective attention and conflict monitoring, the N400 component has been reported to be involved for individuals who have high anxiety (Badzakova-Trajkov et al., 2009; Osinsky et al., 2010). Therefore, although limited, findings from neuroscientific studies suggest that anxiety is involved in erroneous conflict monitoring activities that require updating.



3. Relationship Between Working Memory and Anxiety

The interplay between working memory and anxiety has been researched over the years using a wide variety of working memory tasks. Although the findings from literature lead to a general understanding of how anxiety impacts the working memory, the results are inconsistent. Anxiety has been shown to impair cognitive performance and working memory (Maloney et al, 2016; Moran, 2016). At the same time, cognitive impairment has also been reported to enhance anxiety levels (Petkus et al., 2017). Some studies have indicated that individuals with high trait anxiety have slower reaction times in comparison to those with low trait anxiety (Ladouceur et al. 2009; Elliman et al. 1997). Additionally, there are contradictory views by researchers who have used reading span, word recall, and recognition tasks that resulted in improved working memory despite participants having high trait anxiety (Ferrari and Balconi 2011; Walkenhorst and Crowe 2009; Terry and Burns 2001; Sorg and Whitney 1992). Another study reported that trait anxiety does not influence the working memory (Salthouse, 2012). The findings from previous studies imply a rather mixed opinion on how anxiety affects working memory. Nevertheless, the majority of prior research reports the impairment of working memory under anxiety and stress. The section below summarizes findings from previous studies categorically.

3.1 Anxiety and Working Memory Performance

Moran (2016) conducted a meta-analysis on the relationship between WM performance and anxiety. The result of the analysis on 177 correlative studies indicated anxiety and working memory performance share an inverse relation where higher anxiety dictates lower WM performance. The studies analyzed in the metaanalysis entailed different types of anxiety such as trait, state, clinical, and non-clinical and the tasks tanger from simple span, n-back, complex span spatial, visual, and phonological. The limitation of the meta analysis was that it provides only a general idea on the relationship between anxiety and working memory performance. The task specific and anxiety specific effects have not been considered. There is a lack of research that examines the relationship between state anxiety and a working memory domain.

3.2 Stress and Working Memory

Stress is often considered to be a component of anxiety and also as a separate emotional state that is known to overlap with anxiety. The difference between stress and anxiety is that stress consists of both proactive and avoidance responses, whereas anxiety can be experienced without a neuroendocrine cascade (Miller & O'Callaghan, 2002). The overlap between stress and anxiety is observed in their cognitive effects. When under stress, attentional resources decrease as they are engaged in the potential threat similar to that of anxiety (Klein & Boals, 2001). This effect of stress is mostly observed during high stress levels whereas moderate stress also has the tendency to enhance cognitive performance (Lewis, Nikolova, Chang, & Weekes, 2008). However, research indicates that chronic stress has a negative impact on learning and neuroplasticity. A within-person variability on daily stressors was studied by (2006) among older and young adults on six occasions. The findings of the study showed that the older group had experienced negative effects of high stress level on attention tasks whereas the response of both groups were variable in updating tasks. This finding is supported by the attention depletion hypothesis according to which everyday stressors can deplete attentional and working memory resources.

3.3 Anxiety and Working Memory Updating

Anxiety influences several components of the working memory which leads to impairment in cognitive functioning. A research study by Brose, Schmiedek et al. (2012) investigated how negative stress and anxiety affects young adults in terms of working memory. Under anxious and stressed conditions, the participants performed poorly in the working memory tasks, displayed lack of motivation, and had reduced attentional control. Based on the earlier review, it has been established that a majority of studies indicate that anxiety impairs working memory and cognitive functioning. This raises two important questions on why and how this happens.

How anxiety impairs cognitive abilities and impacts working memory could be attributed to attention bias. Attention bias and limited control over worry are characteristic features of anxiety. These features are known to have a negative impact on cognitive performance (Maloney et al., 2014) such as disruption of working memory (Moran, 2016). In the context of anxiety and impairment of the working memory, the attentional control theory (ACT) is particularly relevant. Eysenck et al. (2007) proposed the attentional control theory which states that cognitive performance is disrupted by state anxiety because the stimulus-driven attentional system is emphasized. The next question on why the effect of anxiety occurs could be explained with brains regions.

Brodmann's areas 46 and 9 have been consistently indicated to be involved in executive control processes (Miller & Cohen, 2001). These brain regions have also been associated with the performance of tasks conducted with working memory manipulation. D'Esposito et al. (1999) reviewed functional magnetic resource image (fMRI) studies on delayed-response tasks designed to assess the prefrontal cortex (PFC) regions which are involved in working memory processes. The study reported that the lateral prefrontal cortex is involved in manipulating and maintaining information, encoding information, and response processes. In a recent fMRI study conducted by Basten et al. (2011), a strong activation was found in the right dorsolateral PFC among individuals who had high anxiety, especially in the inhibitory control exertion needed in the Stroop task. The left dorsolateral PFC, however, has been shown to have a weaker activation compared to that of the right during exertion of inhibition while individuals performed a letter search task ( Bishop, 2009). 

3.4 Trait Anxiety and Working Memory

A growing body of research has highlighted the possibility of functional neurological differences between low anxiety and high anxiety especially while performing tasks that require updating. Basten, Stelzel, & Fiebach (2012) conducted a study to assess whether trait anxiety impairs processing efficiency of manipulation processes and information maintenance. They predicted that working memory manipulation would be impaired due to anxiety and have a negative correlation with neural efficiency during working memory manipulation tasks. The results showed that when high-anxiety participants performed tasks that required attentional control due working memory manipulation, their performance was impaired. Mere maintenance of information was not affected but goal-directed information maintenance and processing were affected, thus suggesting high trait anxiety has a negative impact on working memory. Additionally, the authors also found that the left inferior frontal sulcus and dorsolateral prefrontal cortex which are associated with attention and goal directed control were activated during tasks that required updating. Such differences in working memory tasks between individuals with low anxiety and high anxiety were also observed by Fales et al. (2008). The authors reported neurological differences among the two groups when they performed tasks requiring working memory updating. Although most research works report consistent findings about a negative correlation between high trait anxiety and working memory components, Berggren and Derakshan (2013) suggested that anxiety is associated with working memory updating only under stressful situations which indicates that state anxiety may also be involved. The authors further explained that there is a lack of research that examined the relationship between anxiety and updating using complex tasks such as n-back tasks.



4. Limitations of Prior Research

The predictions of attentional control theory are based on prior studies which involve tasks on storage-oriented working memory tasks (eg., simple or complex span tasks). Such storage oriented working memory tasks do not require updating since there is no requirement of removing irrelevant information with relevant information within a trial. Next, existing research on working memory span tasks have mostly used accuracy rates but not reaction times. In n-back tasks, as the complexity increases, reaction time increases and accuracy decreases (Schmidt et al., 2009; Miller et al., 2009; Harvey et al., 2005 ). Therefore, there is a negative correlation between accuracy and reaction time. The processing efficiency hypothesis of the attentional control theory states that anxiety impacts processing efficiency more measured by reaction time than the effectiveness of performance measured by accuracy rate (Eysenck et al., 2007). Majority of research studies on working memory and anxiety have focused on its storage component which has led to the underestimation of the impact of anxiety on updating (Gustavson & Miyake, 2016). Consequently, there is a lack of studies that focus on research on the relationship between working memory updating and the anxiety through reaction time and accuracy measures.

Previous research studies that have examined the interplay between anxiety and working memory measures have not delved into the subcomponents of anxiety. Most studies have focused on test anxiety or trait anxiety (Visu-Petra et al., 2011; Calvo & Eysenck, 1996; Calvo et al., 1992). The impact of high trait anxiety on working memory tasks has contradictory results due to varying research constructs. Focusing on a single measure of anxiety such as trait anxiety or trait anxiety leads to the underestimation and overlooking of state anxiety. Individuals with high trait anxiety are more likely to have a higher state anxiety compared to individuals with low trait anxiety (Horikawa & Yagi, 2012). Hence, the impairment of cognitive resources and working memory measures in tasks requiring updating would likely be more prominent when performed by individuals with high state and trait anxiety levels. Looking from this point of view, it is necessary to include measures of anxiety apart from trait anxiety to investigate its potential in impacting the working memory. Therefore, these considerations indicate that the relationship between working memory updating and anxiety is not well known, especially in terms of reaction time and accuracy rate. Moreover, the lack of studies that have used working memory updating tasks also potentiate the need to investigate this relationship. Therefore, the present study aims to narrow the literature gap. The findings of the study would be significant in understanding how state anxiety influences updating in terms of accuracy and reaction rate with the help of working memory tasks that require updating.



5. Methods

5.1 Participants

The sample size was determined by running a power analysis based on findings from previous studies which are similar to the present study. Moran (2016) conducted a meta-analysis of 177 studies with 22,061 and reported a correlation coefficient of .33 between anxiety and working memory measures. A meta-analysis of correlation between anxious population and attentional bias towards threat was reported to be 0.45 (Bar-Haim et al., 2007) which indicates a medium effect. Thus, to obtain a power of 0.80 with a = 0.05, based on correlation that ranges from .33 to .45, the sample size should be 99 to 175 (Cohen, 1988).

Our sample consisted of 18 (10 female) participants were recruited where 8 were in the control group (M = 24.25 years; SD= 1.38) and 10 were in the experimental group (M = 23.3 years; SD = 2.75). All participants were Russian natives who were undergraduate or postgraduate students. Our initial plan was to recruit at least 40 participants, however, the current complex circumstances have disallowed so.

5.2 Materials

Mood manipulation techniques were employed through various methods elaborated below and anxiety measures were implemented through a standard scale. Both control and experimental groups were subjected to the same mood manipulation techniques but the stimuli was different.

Anxiety measure.

STAI-40.

State-Trait Anxiety Inventory (STAI), developed by Spielberger et al. (1970) consists of two forms Y1 and Y2 with each form corresponding to state anxiety and trait anxiety. Each form consists of 20 items that to be marked on a 4-point Likert scale (not at all up to very much ). The items on state anxiety (eg., I feel calm, I feel secure, I am worried ) assess the current level of anxiety a person is feeling whereas the items on trait anxiety (eg., I feel like a failure, I am happy, I have disturbing thoughts ) assess how a person generally feels. There are 19 anxiety-absent items and 21 anxiety-present items. The range of score is 2- to 80 where a threshold score of 38 is the highest score of low anxiety, a score of 39 to 40 suggests clinically significant symptoms of state anxiety (Knight, Waal-Manning, & Spears, 1983) whereas a score of 54 to 55 is the cut-off for older adults (Kvaal et al., 2005). Since subjects in the present study were of ages 18 - 30 years old (M = 24.5 years), 39 is considered the cut-off for high anxiety. Regarding reliability of STAI, various studies have reported its reliability over the years indicating high internal consistency among its items. Quek et al. (2004) reported reliability of Cronbach's alpha (б) of 0.86 and Lim et al. (2009) reported its reliability with б = 0.97. Overall, the reports on the reliability of STAI have demonstrated its internal consistency. Therefore, STAI-40 was chosen to measure the baseline anxiety of participants in the present study.

STAI-6.

STAI-6 entails six items from the original T form of the State-Trait Anxiety Inventory where each item corresponds to state anxiety (Spielberger et al., 1983). Among the six items in STAI-6 , three items are anxiety-present (tense, worried, upset) whereas three are anxiety-absent ( content, calm, relaxed ). The items are presented as statements (eg., I feel worried ) and each item is required to be rated on a 4-point Likert scale (not at all up to very much ). The reliability of STAI-6 has been verified by Marteau and Bekker (1992) where they reported Cronbach's alpha of 0.82 and later a reliability of б = 0.79 was reported by Tluczek et al. (2009). Most recently, Lukasik et al., (2019) also reported a reliability of б = 0.82 of STAI-6 in their study on the relationship of working memory performance with stress and anxiety. The reliability of STAI-6 has been demonstrated in various studies and was thus chosen for the present study.

Mood manipulation.

Autobiographical recall.

Autobiographical recall is a mood manipulation technique where researchers ask participants to think and/or write about an experience from their past when they experienced a certain emotion such as happiness or fear. When an experience associated with anger is recalled, a relief of anger is usually felt (Siedlecka, Capper, & Denson, 2015). In the case of an experience associated with fear and anxiety, it may result in an increase in blood pressure, heart rate, skin conductance, and decreases in respiratory rate and respiratory period (Prkachin et al., 1999). In the case of control groups, participants are asked to think and/or write about an ordinary event (Siedlecka, Capper, & Denson, 2015). Recalling ordinary events that do not have emotion-eliciting impact would retina neutral mood.

Visual stimuli.

Visual stimuli such as videos or static images are often used in experiments to evoke specific emotions. To induce anger, an image display of domestic violence and mistreatment have been reported to produce a significant effect (Lobbestael, Arntz, & Wiers, 2008). Monologues, images, cartoons, and film segments with humorous content induces subjective happiness (Lench et al., 2011). Fearful visual stimuli such as films or images have been shown to induce psychological responses that are consistent with fear such as anxiety. Such response is characterized by increased heart rate and startle response (Sarlo et al., 2005). For control groups the visual stimuli chosen must lack emotion evoking content such as clips from nature documentaries (Springer, Rosas, McGetrick, & Bowers, 2007).

Music.

Listening to music through different types of auditory input such as melody or tempo causes the activation of affect (Krumhansl, 2002). Harmonies, fast tempo, and tunes with a dance like rhythm have been shown to induce happiness (Krumhansl, 2002). Listening to fearful music can induce feelings of anxiety (Krumhansl, 1997) and increase blood pressure, heart rate, and the rate of breathing (Etzel et al., 2006). Regarding control groups, either they do not listen to any music at all or auditory stimuli that induce a different emotion than the target emotion is chosen (Sharman & Dingle, 2015).

5.3 Tasks

Affective 2-back task

An affective 2-back task was developed to measure updating of information using pictures from the EU-Emotion Stimulus Set (O'Reilly et al., 2015; O'Reilly et al., 2012) as stimuli. The stimulus set entails photographs of actors of all ages starting from children to adults and older adults who are displaying a wide range of mental and emotional states. Six images were selected from the happiness and excitement categories representing the positive stimuli and six images were chosen from the fear and anger categories representing negative stimuli. All the pictures were converted to grayscale images from color images to prevent attention shift to the colors instead of the facial expression of the actors. Every photo was displayed on the screen for two seconds that was followed by an inter-trial interval of another two seconds. The participant had to press “P” if the picture on display had the same valence as the one displayed two steps back otherwise “Q” had to be pressed if the match condition did not fulfill. The task entailed two blocks that were presented randomly across the participants. Each block contained 42 faces where 40 responses were analyzed and there were 21 positive positive faces and 21 negative faces. anxiety psychological avoidance memory

Non-affective 2-back task.

The non-affective 2-back task was the classic 2-back task with letters as stimuli. It contained two blocks where each block had 42 items where one item was displayed on the screen for two seconds followed by an inter-trial interval of another two seconds. The participant had to press “P” if the letter on display had the same status of either vowel or consonant as the one displayed two steps back otherwise “Q” had to be pressed if the match condition did not fulfill.



5.4 Procedure

The participants were informed during recruitment that the study was investigating emotional information and gave their written consent. The procedure of the experiment is presented in Figure 1. In the beginning of the experiment, the participants were asked to fill the STAI-40 questionnaire. Next, there was a training session for 2-back affective and non-affective versions. After the training, the first mood manipulation technique, autobiographical recall, was applied where the participants had to write down the recalled incident. The experimental group was asked to recall an incident where they felt fear and anxiety while the control group was asked to recall an ordinary day when nothing extraordinary or memorable occured. The paper used for writing was destroyed after the experiment was over. This was followed by the first task which was either affective 2-back or non-affective 2-back task. After the task was completed, the participants had to fill a STAI-6 questionnaire. This was followed by the second mood manipulation technique where a visual stimuli was presented. The experimental group was shown a fearful, short video clip from a movie where a man appeared to attempt to attack a young boy. The clip ended before the movie clip showed whether the boy was alive after the attack, leaving the viewer in a state of worry. The control group was shown a video clip of a math tutorial on area by integration that was devoid of emotion evoking stimuli. After the second mood induction, the participants had to complete the second task which was either non- affective 2-back or affective 2-back task. The experiment was concluded with the filling of STAI-6 to measure state anxiety. During the tasks, a background music was being played. A neutral music was chosen for the control group and a fearful music was chosen for the experimental group. Subsequently, the participants were debriefed on the actual goal of the experiment. The experiments were conducted anonymously where only the age and sex of the paricapnts were recorded. The 2-back tasks were created using PsychoPy (Peirce et al., 2019) and lasted for 40 minutes.



Figure 1. Overview of experimental procedure



6. Data Analysis and Results

6.1 STAI-40 and STAI-6

Paired-sample independent t-test was run on the STAI-scores to analyze the baseline anxiety and the effect of anxiety after the completion of tasks of both experimental and control groups. Mean state and trait anxiety scores presented in Table 1 show that the baseline state and trait anxiety of the control group (M = 31.4, SD = 16.945 and M = 37.5, SD = 20.592, respectively) was less than that of the experimental group (M = 36.7, SD = 5.945 and M = 43.2, SD = 4.962, respectively). The scores indicate low anxiety except for mean trait anxiety score of the experimental group (M = 43.2, SD = 4.962) which is greater than 40 indicates high level of anxiety.

Table 1. Mean scores and standard deviation of STAI-40 (state and trait), STAI-6 after completion of each task ( 2-back affective and 2-back non-affective).

	Mean of STAI-40 (State)	Mean of STAI-40 (Trait)	Mean of STAI-6 (after Task 1)	Mean of STAI-6 (after Task 2)
Control Group Experimental Group	31.4 (SD = 16.945)	37.5 (SD = 20.592)	12.875 (SD =2.416)	13.875(SD=2.255)
	36.7 (SD =5.945 )	43.2 (SD = 4.962 )	13.6 (SD =2.412 )	15.1 (SD=3.281 )

The 10 participants in the experimental group compared to the 8 participants in the control group did not demonstrate a significant difference in their baseline anxiety and state anxiety after completing the 2-back affective and non-affective tasks (Table 2). Although there was no significant effect (p= .536) for state anxiety after completion of the first task, the mean state anxiety (M =13.6, SD = 2.412) for the experimental group was higher than that of the control group (M = 12.875, SD = 2.416). Similarly, there was a non-significant effect (p = .366) of state anxiety after completing the second task, however, the mean STAI-6 score was higher in the experimental group (M =15.1, SD = 3.281) than the control group (M = 13.875, SD = 2.295). Medium effect size exists between the state anxiety (d = .446) and trait anxiety (d = .446) of control groups and experimental groups. A small effect (d=.300) is seen in the state anxiety after completion of the first task between the two groups whereas the effect is medium after completion of the second task (d=.439).

Table 2. Results from paired-sample t-test and Cohen's D of baseline anxiety (state and trait) and state anxiety after task completion between experimental and control groups.

Test		p-value	Cohen's D
STAI - 40	State Anxiety Trait Anxiety	.414	.446
		.370	.463
STAI - 6	Task 1 Task 2	.536	.300
		.366	.439

6.2 Affective and Non-Affective 2-back Tasks

The mean accuracy rates and mean reaction time of the correct responses of both versions of the 2-back tasks, affective and non-affective were used as indices for working memory updating. The mean accuracy of the experimental group (M = .378, SD = .077) in the non-affective task (Figure 2) was lower than the control group (M = .422, SD = .034) (Table 3). The same trend was observed in the affective 2-back task (Figure 3) where the experimental group had a lower mean accuracy rate (M = .376, SD = .106) than that of the control group (M = .415, SD = .103). This finding is in support of the hypothesis which states that state anxiety leads to impairment of updating in the working memory since updating is essential in n-back tasks.



Figure 2. The accuracy of the 2-back non-affective task of the experimental group is lower than that of the control group, as indicated by the interpolation line.

...