Memory and language impairments are associated with anxiety disorder severity in childhood.

Introduction Children with anxiety disorders have been suggested to possess deficits in verbal fluency, shifting and attention, with inconsistent results regarding working memory and its subcomponents. This study extends previous findings by analyzing the performance of children with anxiety disorders in a wide range of neuropsychological functions. Methods We evaluated 54 children with a primary diagnosis of an anxiety disorder according to diagnostic criteria from the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) using subtests of a neuropsychological battery. The severity of anxiety disorders was assessed using the Pediatric Anxiety Rating Scale (PARS). We calculated the frequency of neuropsychological impairments (-1.5 standard deviation of the normative sample). Comparisons between groups were performed based on the severity of anxiety symptoms, as well as in the presence of one vs. more diagnoses of anxiety disorder. Results We found higher impairment in visuospatial working memory (23.1%), semantic memory (27.8%), oral language (35.4%) and word writing (44.4%) in anxious children. Moreover, children with higher anxiety severity presented lower performance in visuospatial working memory, inferential processing, word reading, writing comprehension, copied writing, and semantic verbal fluency (d = 0.49 to 0.96 [Cohen's d]). The higher the number of anxiety diagnoses, the lower the performance in episodic memory and oral and written language (d = 0.56 to 0.77). Conclusion Our data suggested the presence of memory (visuospatial working memory and semantic memory) and language deficits (oral and writing) in some children with an anxiety disorder. Severity and number of anxiety diagnoses were associated with lower performance in memory and language domains in childhood.


Introduction
Pediatric anxiety disorders are among the most common mental disorders in Brazil. In 2015, one study reported that between 4.2 and 9.4% of the children of four Brazilian regions were affected by these disorders. 1 The conditions often have a chronic course and are frequently associated with considerable impairment and dysfunction. 2  Most of the research on anxiety disorders has investigated the role of anxious symptoms on cognitive functions like attention and executive functions. 3,4 According to some authors, attention can be interpreted as being divided into three main dissociable networks: 1) executive attention (ability to solve and monitor conflicts in process competing for stimuli and responses); 2) alerting (controlling keenness to respond to novel stimuli); and 3) orienting attention that sets up sensory information. 5 Attention orienting towards threats has been one of the most replicable results for anxiety disorders. In addition to the role of symptoms in attention orienting, positive results have also been found for deficits in some executive function components, such as phonemic verbal fluency, 6,7 executive attention, 8 and working memory. [9][10][11] Comprehensive investigations of distinct aspects of memory (i.e., semantic and episodic verbal memory) and language (i.e., oral and written) are scarce, with some exceptions. 12,13 It is also known that most researches did not evaluate the role of severity when investigating neuropsychological impairments in children with anxiety disorders. Some studies have shown, for example, that severity is very important when investigating the role of threat bias in anxiety disorders, as well as when investigating deficits in distinct aspects of memory. 13,14 Neuropsychological studies have described cognitive dysfunctions in children and adolescents with anxiety disorders, but their results are inconsistent.
In a study evaluating attention, verbal episodic memory, working memory, visuoconstructive skills, and executive functions, no differences were found in patients with anxiety disorders as compared to a non-anxious control group, except for the digit span backward task. 12 Furthermore, the group with mild anxiety disorder presented higher performance than the control group in this task. 12 Another study, using the same community sample, showed that youths with anxiety had higher deficits in verbal fluency when compared with the non-anxious group. 6 This finding was replicated and extended through young children, showing that verbal fluency is consistently associated with severity of anxiety disorders, regardless of the presence of attention-deficit/hyperactivity disorder (ADHD) symptoms. 7 Poor working memory performance is one of the neuropsychological functions most frequently associated with pediatric anxiety disorders. 9,10 Working memory is usually described as a framework for a series of interactive processes that comprise temporary storage and the manipulation of information. 15  Anxiety disorders in children and adolescents are also associated with low language skills and low shifting performance, possibly mediated by decreased attention, decreased short-term memory or working memory. 11 However, the study of Toren et al. 11 did not find an association between anxiety and nonverbal processes. willing to participate in a randomized clinical trial for psychological treatment. 22 Participants were recruited via mass media advertising that asked for children who had difficulty/fear of being separated from parents or liaison figures; excessive concern with everything; or excessive shyness) and were first screened by a telephone interview. After the telephone interview,   26 We adapted some of the questions to reflect changes made in the DSM-5. The K-SADS-PL has been adapted to Brazilian Portuguese and presented good psychometric properties. 27

Pediatric Anxiety Rating Scale (PARS)
The PARS is a clinician-rated measurement of anxiety severity. This instrument rates anxiety severity, frequency, distress, avoidance, and interference in daily functioning, in accordance with standardized methods. 28 We used the scores on the 50 symptoms evaluated as were performed using Mann-Whitney's U test. To deal with alpha inflation due to multiple comparisons, the false discovery correction rate was applied (FDT). 32 We used the equation r = z/√N to calculate the effect size of the differences between groups. 33 In the formula, z is the z distribution and N is the study sample size. In order to compare the calculated effect size with other effect sizes (for example Cohen's d), the following formula was used: d = 2r/√(1-r2). 33 Effect sizes were classified as small (0.2), moderate (0.5) and large (0.8), according to Cohen's guidelines. 34 We adopted a 5% chance of significance. All analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 24.0. Table 1 shows the frequency of children diagnosed with anxiety disorder scoring below the mean (z < 1.5

Results
SD of the normative sample) in the cognitive domains of the neuropsychological battery. Children had difficulties in memory (visuospatial working memory and semantic memory) and oral and written language subtests. In the word writing subtest, about 44% (24 children) presented impaired performance.
Our data showed statistically different scores in working memory and oral language according to severity of the anxiety disorder in our sample. More specifically, Table 2 shows that, when compared to children with low anxiety severity, children with high severity had lower medians in visuospatial working memory, oral language (comprehension and processing of inferences), written language (word reading, comprehension, spontaneous writing and copied writing) and semantic verbal fluency. In this latter function, the difference reached a large effect size (d = 0.96). Table 3 shows the neuropsychological subtests in which there were statistically significant differences between the groups according to number of anxiety disorders. Children with two or three comorbid diagnoses of anxiety disorders performed poorly in six domains as compared to the group with only one diagnosis of anxiety disorder, with moderate effect sizes. There were no significant differences in the other neuropsychological subtests (Table S1, available as online-only supplementary material).

1) Visual attention: Image cancellation -Requires focused attention on visual mode and ability of inhibition.
The final score is the number of targets correctly canceled. 2) Auditory attention: Digit span forward -Children are told a sequence of digits (2 to 5 items) and they have to repeat them in the exact same order.

Working memory
Phonological and central executive: 1) Digit span backward -Children are told a sequence of digits (2 to 5 items) and they have to repeat them in the inverse order. 2) Pseudoword span -Children are told a sequence of pseudowords and they have to repeat them. The number of stimuli progressively increases from 1 to 4 items.
Visuospatial operational: 3) Working memory (operational) visuospatial in the reverse order -The examiner points out progressively longer sequences of stimuli (i.e., squares that were randomly arranged on a blank sheet), and the child is asked to repeat pointing out the stimuli in the inverted presentation order immediately after the model was presented by the examiner.
Verbal and visual episodic-semantic memory 4) Immediate and late recall (words) -Consists of the oral recall of 9 words. 5) Immediate recall (pictures) -Consists of the oral recall of 9 pictures. 6) Semantic memory -Consists of 4 questions that access the child's prior knowledge (long-term memory).
Phonological awareness: 2) Rhyme -Children are presented with 3 words that are named by the examiner. Children are asked to identify the 2 of them that rhyme (or that "are sounding the same"). The task consists of 2 practice items and 4 test trials (maximum score: 4 points). 3) Phonemic subtraction task -Requires the participants to omit the initial phoneme (or the final phoneme) from a nonword. Children are presented a short nonword orally and asked to state the nonword without the first (or last) sound.
Oral comprehension: 4) Oral comprehension -Words and phrases are presented orally. Children should point to the drawing corresponding to what is being said. There are 3 options to choose: 1 correct and 2 distracting.
Inferential processing: 5) Processing of inferences -Metaphors and proverbs are presented orally. Children should explain their meaning.
Written language 6) Reading aloud syllables, words, and pseudowords -The stimuli are 6 syllables, 6 real words (regular and irregular; frequent and infrequent; short and long) and 5 pseudowords. 7) Writing comprehension -Five words and phrases are presented one by one in writing. The child should read silently and then point at the target figure. Two distractors are presented for each word or phrase. 8) Writing words and pseudowords -Stimuli are dictated by the examiner, one at a time. Only after the end of the word/nonword can the child start writing. Words are dictated in the usual pronunciation of the region. The stimuli are 14 real words (with 4 monosyllables) and 5 pseudowords. 9) Spontaneous writing -Involves writing a complete sentence. 10) Writing copied -Involves copying a complete sentence.

Executive functions
1) Semantic verbal fluency -Children are asked to say animal names in 1 minute.
2) Inhibitory control -By the auditory go/no go task, in which 0 to 9 digits are presented to the child at a rate of 1 item per second, and the child has to respond "yes" each time the child listens to one digit, except for the digit 8, for which the child should remain silent.
For more details about the test, description and rating of the tasks, see Salles et al. 30   showing that these deficits are also present in semantic verbal fluency and are also related to the severity of anxiety symptoms in children with a diagnosis of anxiety disorder. Verbal fluency is widely considered a measure of executive functioning. 42,43 Other studies also show a dose-response relationship between anxiety severity and other measures of executive function, such as conflict scores in the attention network. 8,44,45 These findings raised questions on whether those deficits are specific to anxiety disorders, or if they are related to the "p-factor", i.e., the overall severity of psychopathology across disorders, shared between both internalizing and externalizing disorders, as suggested by the referred studies.

Conclusions
The present study reinforces previous findings that cognitive functions (e.g., memory and language) might be compromised in anxiety disorders in children, even in situations where there is no emotional context (e.g.., threat). 6,12 Interventions targeting cognitive mechanisms at early age may be studied to minimize future problems related to anxiety disorders.