Visuospatial and Mathematical Dysfunction in Major Depressive Disorder and/or Panic Disorder: A Study of Parietal Functioning

Abstract

The parietal cortex is critical for several different cognitive functions, including visuospatial processing and mathematical abilities. There is strong evidence indicating parietal dysfunction in depression. However, it is less clear whether anxiety is associated with parietal dysfunction, and whether comorbid depression and anxiety is associated with greater impairment. The present study compared participants with major depression (MDD), panic disorder (PD), comorbid MDD/PD, and controls on neuropsychological measures of visuospatial processing, Judgment of Line Orientation (JLO), and mathematical abilities, Wide Range Achievement Arithmetic (WRAT-Arithmetic). Only comorbid MDD/PD was associated with decreased performance on JLO, whereas all psychopathological groups exhibited comparably decreased performance on WRAT-Arithmetic. Furthermore, the results were not accounted for by other comorbid disorders, medication use, or psychopathology severity. The present study suggests comorbid depression and anxious arousal is associated with impairment in visuospatial processing and provides novel evidence indicating mathematical deficits across depression and/or anxiety. Implications for understanding parietal dysfunction in internalizing psychopathology are discussed.

Introduction

The interplay between cognition and emotion is critical to various forms of psychopathology. The parietal cortex is one region that is important for emotional arousal (Heller, 1993; Heller, Nitschke, Etienne, & Miller, 1997), and there is evidence that variations in arousal can interfere with cognitive functions that share similar neural substrates. For example, Shackman et al. (2006) found that threat-of-shock increased physiological arousal and selectively disrupted cognitive processes (e.g., visuospatial working memory) that are asymmetrically associated with the right frontal and parietal regions. The authors in that article suggested that the task-irrelevant arousal competed for limited resources with ongoing cognitive functions that also involved the right parietal region. Therefore, conditions characterized by aberrant emotional arousal may be particularly vulnerable to disruptions in parietal functions.

There is a robust literature indicating parietal dysfunction in depression. Indeed, several studies have tested the hypothesis of low emotional arousal in depression using parietal electroencephalography (EEG) asymmetry and shown that a reduced relative right parietal EEG asymmetry is associated with increased depressive symptoms (Blackhart et al., 2006; Rabe et al., 2005) and differentiates remitted depression from controls (Henriques & Davidson, 1990; Stewart et al., 2011). However, some investigations have failed to find an association between parietal EEG asymmetry and depression (Debener et al., 2000; Henriques & Davidson, 1991; Nitschke et al., 1999; Schaffer et al., 1983). One possible explanation for the mixed findings is unmeasured comorbid anxiety, which is posited to be associated with increased emotional arousal. Indeed, several studies have reported that comorbid depression and anxiety is associated with a greater relative right parietal EEG asymmetry (Bruder et al., 1997; Kentgen et al., 2000; Mathersul et al., 2008; Metzger et al., 2004).

Other studies have examined parietal dysfunction in depression and anxiety using neuropsychological measures of perceptual asymmetry. For example, several investigations have used the chimeric faces task (Levy et al., 1983) to examine hemispatial bias, such that changes in left hemispatial bias indicate changes in right parietal activity (Green et al., 1992). As expected, depression has been associated with a reduced left hemispatial bias (i.e., reduced right parietal activity) and anxiety has been associated with an increased left hemispatial bias (i.e., increased right parietal activity) (Heller et al., 1995; Jaeger et al., 1987; Keller et al., 2000). Perceptual asymmetry has also been examined using dichotic listening tasks, as the difference in performance accuracy to the left and right ears provides information on hemispheric lateralization. Across several studies depression has been associated with a greater right ear advantage (i.e., decreased right temporoparietal activity) and anxiety disorders have been associated with a greater left ear advantage (i.e., increased right temporoparietal activity) (Bruder et al., 1999; Pine et al., 2000; Wexler & Goodman, 1991). Furthermore, in one of the most thorough dichotic listening investigations using DSM-determined diagnoses, Bruder and colleagues (2004) found that anxiety with or without a comorbid depressive disorder was associated with a smaller right ear advantage (consistent with decreased left parietal activity) relative to depression only and controls.

Despite the supporting evidence, there are still limitations to the literature on parietal functioning in depression and anxiety. First, several theorists have distinguished between two subtypes of anxiety disorders or dimensions – anxious apprehension (e.g., generalized anxiety disorder) and anxious arousal (e.g., panic disorder), and suggested that the former is associated with increased left frontal regions while the latter is associated with increased right parietal regions (Heller & Nitschke, 1998; Heller et al., 1997). Few studies have examined the role of comorbid anxious arousal and those that did largely relied on self-report symptoms in non-clinical samples.

Second, the literature on depression and anxiety and parietal functioning has primarily focused on measures of EEG and perceptual asymmetry. The parietal cortex is critical for a number of different cognitive functions, including attention, executive functioning, mathematical abilities, and visuospatial processing (Collette et al., 2005; Constantinidis, Bucci, & Rugg, 2013; Kolb & Whishaw, 1996; Malhotra, Coulthard, & Husain, 2009; Rivera, Reiss, Eckert, & Menon, 2005). There is also strong evidence for hemispheric lateralization for several of these functions. For example, right parietal regions have been associated with visuospatial processing in neuroimaging (Clements et al., 2006; Deutsch et al., 1988) and lesion studies (Tranel et al., 2009; Treccani et al., 2005). Conversely, left parietal regions have been associated with mathematical abilities in neuroimaging (Chochon et al., 1999), lesion (Boller & Grafman, 1983; Jackson & Warrington, 1986), and cortical stimulation studies (Whalen et al., 1997). While the right parietal-visuospatial association has been more robustly found than the left parietal-math association, acalculia (a disorder in math abilities) has been consistently reported in patients suffering lesions in left parietal regions (particularly the angular gyrus), relative to other regions (Mayer et al., 2003). Additionally, it is important to highlight that these (and other) cognitive functions involve additional regions outside the left and right parietal cortices (Ng et al., 2000; Pinel & Dehaene, 2009), respectively, and are affected by dysfunction in other brain areas. To date, a small number of studies have found an association between depression and impaired visuospatial processing (Coello et al., 1990; Paradiso et al., 2001) and mathematical performance (Lagace et al., 2003). However, it is unclear whether anxiety disorders characterized by heightened emotional arousal (e.g., panic disorder, a disorder often associated with heightened anxious arousal, McNally, 2002) are also associated with these cognitive deficits, and whether comorbid depression and anxiety is associated with greater impairment.

The present study aimed to address these limitations and compared participants with DSM-IV diagnosed: 1) major depressive disorder (MDD) without a lifetime history of an anxiety disorder, 2) panic disorder (PD) without a lifetime history of depression, 3) comorbid MDD and PD, and 4) healthy controls on two separate tasks that measured visuospatial processing, Benton’s Judgment of Line Orientation (JLO) test (Benton et al., 1978), and mathematical abilities, the Wide Range Achievement – 3 Arithmetic (WRAT-Arithmetic) subtest (Wilkinson, 1993). This study was designed to extend the extant EEG and perceptual asymmetry literature on parietal functioning in depression in a novel way—by using alternative neuropsychological measures that primarily involve parietal cognitive functions (visuospatial and mathematical abilities) and a well-defined clinical sample that allowed for comparison of the independent and interactive effects of depression and anxious arousal on parietal functioning.

The present study had two primary hypotheses. First, the study hypothesized that MDD (with or without comorbid PD) would be associated with decreased performance on the JLO. However, the few studies that have examined comorbid depression and anxiety have generally found abnormal right parietal activity in comorbid relative to ‘pure’ depressed groups (Bruder et al., 1997; Kentgen et al., 2000; Mathersul et al., 2008; Metzger et al., 2004). It is therefore possible that only comorbid MDD and PD will be associated with decreased performance on JLO. Second, only a few studies have specifically examined left parietal functioning in depression and anxiety and in those that did, the conditions were associated with impaired performance (Bruder et al., 2004; Lagace et al., 2003). Therefore, we tentatively hypothesized that both MDD and/or PD would be associated decreased performance on WRAT-Arithmetic.

Method

Participants

The sample consisted of 39 individuals with current MDD and no current or lifetime history of an anxiety disorder (MDD-only), 27 individuals with current PD and no current or lifetime history of MDD or dysthymia (PD-only), 54 individuals with current MDD and current PD (comorbid MDD/PD), and 65 controls with no lifetime history of an Axis I disorder.¹ Diagnosis was examined as two 2-level factors, Depression Status (Present vs. Absent) and Panic Status (Present vs. Absent), instead of one 4-level factor in order to examine main effects and interactions of MDD and PD on neuropsychological performance. Diagnoses were made via the Structured Clinical Interview for DSM-IV (SCID; First et al., 1996) and based on criteria defined by the Diagnostic and Statistical Manual of Mental Disorders – Fourth Edition (DSM-IV; American Psychiatric Association, 1994). All interview assessments were conducted by S.A.S. and advanced clinical psychology doctoral students who were trained to criterion by S.A.S. See Shankman et al. (2013) for additional details regarding the sample and diagnostic procedures.

MDD-only participants were required to have no current or past history of an anxiety disorder. PD-only and comorbid MDD/PD participants were allowed to meet criteria for additional current and past anxiety disorders. Both MDD-only and comorbid MDD/PD participants were required to have an age of onset of first depressive disorder (dysthymia or MDD) before 18, because the neuropsychological measures were administered as part of a larger study on early-onset depression (Shankman et al., 2013). Control participants were required to have a 24-item Hamilton Rating Scale for Depression (HRSD; Hamilton, 1960) and Beck Anxiety Inventory (BAI; Beck et al., 1988) score of less than 8. Seven participants (3 MDD-only, 2 PD-only, 1 comorbid MDD/PD, and 1 control) did not complete the BAI.

Exclusion criteria included a lifetime diagnosis of schizophrenia or other psychotic disorder, bipolar disorder, or dementia; inability to read or write English; history of head trauma with loss of consciousness; or being left-handed (as confirmed by the Edinburgh Handedness Inventory, range +20 to +100; Oldfield, 1971). Participants were recruited through advertising in the community (e.g., flyers, Internet postings) and mental health clinics in the greater Chicago area. All participants gave informed consent and were paid for their participation.

Neuropsychological Tasks

Judgment of Line Orientation (JLO)

The JLO test was designed to evaluate visuospatial skills by assessing the ability to judge the orientation and angles of lines in space (Benton et al., 1978). The test consisted of five practice items followed by 30 test items. For each item, a stimulus card was shown at the top of the test booklet and a response card was shown at the bottom of the test booklet. Participants were asked to match a pair of partial-length lines appearing on the stimulus card to two of the 11 numbered full-length lines appearing on the reference card. The 11 full-length lines on the reference card formed a semicircle and were separated by 18-degree intervals. A response was scored as correct only when both lines chosen on the stimulus card matched the lines on the reference card. During the practice trials, participants were corrected if they made a mistake, but received no feedback during the 30 test items. The total number of correct items was out of 30 and there was no time limit to the test. The JLO has two alternate forms (H and V) that consist of the same 30 items presented in slightly different sequences. All participants completed form V. Four participants (2 MDD-only, 1 comorbid MDD/PD, and 1 control) had JLO scores that were greater than 3 standard deviations below the mean and these outliers were excluded from analyses.

Wide Range Achievement Test – 3 (WRAT-3) Arithmetic Subtest

The WRAT-3 Arithmetic subtest measures counting, basic arithmetic, and written computation (Wilkinson, 1993) and consists of oral (15 items) and written (40 items) arithmetic sections. Participants were given 15 minutes to complete as many of the written arithmetic items as possible and the total number of correct items was out of 55. The WRAT-3 Arithmetic subtest has two versions (Blue and Tan). All participants completed the Tan version.

Results

Demographics and Clinical Characteristics

Demographics and clinical characteristics are presented in Table 1. Diagnostic groups were matched on age, education, sex, and race. MDD-only, PD-only, and comorbid MDD/PD participants did not differ on current psychiatric medication use. MDD-only and comorbid MDD/PD participants did not differ on age of onset of first depressive disorder and PD-only and comorbid MDD/PD participants did not differ on age of onset of PD. MDD-only and comorbid MDD/PD participants had greater rates of lifetime alcohol abuse/dependence disorder and lifetime substance abuse/dependence disorder relative to PD-only participants, but did not differ from each other. Comorbid MDD/PD participants had greater rates of lifetime obsessive-compulsive disorder (OCD), post-traumatic stress disorder (PTSD), and social phobia relative to PD-only participants, and PD-only participants had greater rates of lifetime generalized anxiety disorder (GAD) relative to comorbid MDD/PD participants. As expected, control participants had higher global assessment of functioning (GAF) scores relative to MDD-only, PD-only, and comorbid MDD/PD participants; MDD-only and comorbid MDD/PD participants had higher HRSD scores relative to PD-only and control participants; and PD-only and comorbid MDD/PD participants had higher BAI scores relative to control participants.

Table 1.

Demographics and Clinical Characteristics

	Control (n = 65)	MDD-only (n = 39)	PD-only (n = 27)	Comorbid MDD/PD (n = 54)
Demographics
Age (years; SD)	32.6(13.3)	30.7(11.2)	34.1(13.4)	35.9(11.6)
Sex (% female)	59.4%	59.5%	63.0%	69.8%
Race (% Caucasian)	42.2%	51.4%	48.1%	49.1%
Education
Grade 7 to 12 (without graduating high school)	0.0%	0.0%	0.0%	5.7%
Graduated high school or high school equivalent	7.8%	0.0%	3.7%	3.8%
Part college	28.1%	48.6%	18.5%	43.4%
Graduated 2 year college	0.0%	0.0%	14.8%	3.8%
Graduated 4 year college	35.9%	27.0%	44.4%	24.5%
Part graduate/professional school	21.9%	5.4%	14.8%	7.5%
Completed graduate/professional school	6.3%	5.4%	3.7%	11.3%
Clinical Characteristics
GAF (SD)	88.8(7.5)a	53.9(7.1)b	59.0(8.2)b,c	52.2(6.1)b
HRSD (SD)	1.4(1.7)a	24.4(7.9)b,c	8.2(7.0)b	26.7(8.8)b,c
BAI (SD)	1.7(2.0)a	14.3(11.1)b	14.1(9.7)b	21.7(13.2)b,c
Age of onset of first depressive disorder (year)	-	13.6	-	13.4
Age of onset of PD (year)	-	-	20.9	16.0
Lifetime alcohol abuse/dependence disorder	6.3%a	43.2%b,c	25.9%b	52.9%b,c
Lifetime drug abuse/dependence disorder	1.6%a	27.0%b,c	14.8%b	41.4%b,c
Lifetime GAD	-	-	25.9%a	1.9%b
Lifetime OCD	-	-	0.0%b	15.1%a
Lifetime PTSD	-	-	7.4%b	32.1%a
Lifetime social phobia	-	-	7.4%b	35.8%a
Lifetime specific phobia	-	-	22.2%	18.9%
Current psychiatric medications
Any medication	-	27.0%	25.9%	43.4%
SSRI/SNRI	-	18.9%	22.2%	20.8%
Tricyclic/Tetracyclic Antidepressant	-	2.7%	0.0%	5.7%
Atypical Antidepressant	-	0.0%	0.0%	7.5%
Atypical Antipsychotic	-	0.0%	0.0%	5.7%
Benzodiazepine	-	8.1%	7.4%	22.6%
Other	-	8.1%	0.0%	13.2%

Note. BAI = Beck Anxiety Inventory; GAD = Generalized Anxiety Disorder; GAF = Global Assessment of Functioning; HRSD = Hamilton Rating Scale for Depression; MDD = Major Depressive Disorder; OCD = Obsessive-Compulsive Disorder; PD = Panic Disorder; PTSD = Post-Traumatic Stress Disorder; SD = Standard Deviation; SSRI = Selective Serotonin Reuptake Inhibitor; SNRI = Serotonin-Norepinephrine Reuptake Inhibitor.

Demographics and Neuropsychological Performance

Age was negatively associated with JLO at a trend level, r(181) = −.15, p < .06, and WRAT-Arithmetic, r(181) = −.26, p < .001. Education differentiated scores on JLO, F(6, 174) = 4.52, MSE = 73.47, p < .001, η_p² = .14, and WRAT-Arithmetic, F(6, 174) = 6.67, MSE = 138.34, p < .001, η_p² = .19, such that more education was related to better performance. Women (M = 22.72, SD = 4.47) had lower scores than men (M = 25.22, SD = 3.37) on JLO, F(1, 179) = 15.77, MSE = 264.71, p < .001, η_p² = .08, but there were no sex differences on WRAT-Arithmetic, F(1, 179) = 2.17, MSE = 53.08, ns. Finally, Caucasian participants (M = 24.84, SD = 3.33) had higher scores than non-Caucasian participants (M = 22.59, SD = 4.72) on JLO, F(1, 179) = 13.33, MSE = 226.52, p < .001, η_p² = .07, but there were no racial differences on WRAT-Arithmetic, F(1, 179) = 1.10, MSE = 27.03, ns. Therefore, age, education, sex, and race were included as covariates for all analyses involving JLO, and age and education were included as covariates for all analyses involving WRAT-Arithmetic.

Neuropsychological Performance

Diagnostic group differences in neuropsychological performance were examined using analysis of covariance (ANCOVA) with Depression Status (Present vs. Absent) and Panic Status (Present vs. Absent) entered as between-subjects factors and the above demographics entered as covariates.

JLO

Figure 1 (top) displays means (and standard errors) for each diagnostic group on JLO performance. Results indicated a trend level main effect of Depression Status, F(1, 173) = 3.39, MSE = 50.34, p < .07, η_p² = .02, that was qualified by a Depression Status X Panic Status interaction, F(1, 173) = 5.27, MSE = 89.08, p < .05, η_p² = .03. Follow-up analyses indicated that comorbid MDD/PD participants had lower scores on JLO relative to MDD-only, F(1, 84) = 4.87, MSE = 78.82, p < .05, η_p² = .06, PD-only, F(1, 74) = 7.19, MSE = 108.07, p < .01, η_p² =.09, and control participants, F(1, 111) = 4.47, MSE = 66.09, p < .05, η_p² = .04. MDD-only, PD-only, and control participants did not differ on JLO performance (ps > .25).

Figure 1.

Diagnostic group differences on the JLO (top) and WRAT-Arithmetic (bottom). Error bars represent standard error. JLO = Judgment of Line Orientation; MDD = major depressive disorder; PD = panic disorder; WRAT = Wide Range Achievement Test.

Across the entire sample comorbid MDD/PD participants had the lowest GAF scores, and, relative to PD-only participants, had greater rates of lifetime OCD, PTSD, social phobia, alcohol abuse/dependence disorder, and substance abuse/dependence disorder. Therefore, it is possible that the impairment in JLO performance was not due to the comorbidity between MDD and PD, but rather increased psychopathological severity. To address this issue, we conducted an additional analysis that included GAF scores and lifetime alcohol abuse/dependence disorder, substance abuse/dependence disorder, GAD, OCD, PTSD, social phobia, and specific phobia diagnoses as additional covariates. Results again indicated a Depression Status X Panic Status interaction, F(1, 165) = 8.36, MSE = 113.34, p < .01, η_p² = .05, indicating that comorbid MDD/PD participants still demonstrated impaired JLO performance relative to MDD-only, PD-only, and control participants even after controlling for GAF severity and all other comorbid lifetime disorders.

WRAT-Arithmetic

Figure 1 (bottom) displays means (and standard errors) for each diagnostic group on WRAT-Arithmetic performance. Results indicated a main effect of Panic Status, F(1, 175) = 5.12, MSE = 5.12, p < .05, η_p² = .03, that was qualified by a Depression Status X Panic Status interaction, F(1, 175) = 4.38, MSE = 81.64, p < .05, η_p² = .02. Follow-up analyses indicated that MDD-only, PD-only, and comorbid MDD/PD participants had lower scores on WRAT-Arithmetic relative to controls, F(1, 97) = 7.71, MSE = 136.68, p < .01, η_p² = .07; F(1, 87) = 8.81, MSE = 167.85, p < .01, η_p² = .09; F(1, 113) = 8.90, MSE = 161.37, p < .01, η_p² = .07, respectively. MDD-only, PD-only, and comorbid MDD/PD participants did not differ on WRAT-Arithmetic performance (ps > .52).

It is possible that poor performance on WRAT-Arithmetic could have been due to a deficit in psychomotor processing speed and not mathematical abilities given that the test was timed. To address this issue, we also compared diagnostic groups on the total number of WRAT-Arithmetic items that were attempted (i.e., total correct + total incorrect) using an ANCOVA with Depression Status (Present vs. Absent) and Panic Status (Present vs. Absent) entered as between-subjects factors and age and education entered as covariates. Results indicated no main effects of Depression Status, F(1, 175) = 2.19, MSE = 44.21, ns, Panic Status, F(1, 75) = 2.16, MSE = 43.58, ns, or Depression Status X Panic Status interaction, F(1, 175) = 1.22, MSE = 24.64, ns. These results suggest that the poor WRAT-Arithmetic performance by MDD and/or PD participants was not necessarily due to slowed psychomotor processing speed or an inability to attempt a comparable number of test questions.

Medication

Finally, we examined whether medication use differentiated JLO and WRAT-Arithmetic performance. Within the MDD-only, PD-only, and comorbid MDD/PD participants, a one-way ANOVA was conducted with medication status (currently taking psychiatric medication vs. not currently taking psychiatric medication) entered as a between-subjects factor for each performance variable. Results were non-significant for all analyses (ps > .75) indicating that current medication use did not account for the results.

Discussion

The present study examined neuropsychological performance on measures of visuospatial processing (JLO) and mathematical abilities (WRAT-Arithmetic) in a sample of participants with MDD-only, PD-only, comorbid MDD/PD, and controls. There were several noteworthy findings. First, individuals with comorbid MDD/PD exhibited decreased performance on JLO relative to controls and the other two psychopathology groups, and this was not due to greater symptom severity or comorbidity of other psychopathological conditions in the participants with comorbid MDD/PD. Second, individuals in all three psychopathology groups exhibited decreased performance on WRAT-Arithmetic relative to controls, and this was not due to slowed psychomotor processing speed or an inability to attempt a comparable number of test questions. Finally, current psychiatric medication status was not related to JLO or WRAT-Arithmetic performance.

These results provide novel evidence for the presence of cognitive impairment mediated (in part) by parietal regions in depression and anxiety. Specifically, comorbid depression and anxious arousal appear to contribute to visuospatial deficits. Depression has previously been associated with decreased performance on the JLO (Coello et al., 1990; Paradiso et al., 2001); however, these studies did not examine the role of comorbid anxiety. Interestingly, other research has reported visuospatial processing deficits in comorbid depression and anxiety relative to depression only and controls (Suslow et al., 2004). The present finding of mathematical dysfunction in those with depression and/or anxiety is also quite novel. There have only been a limited number of studies that have examined the relationship between mood disorders and mathematical dysfunction (Lagace et al., 2003). Furthermore, we do not know of any study that has examined the association between mathematical abilities and anxious arousal (e.g., PD). Thus, this is some of the first evidence to suggest that MDD and/or PD are associated with comparable levels of mathematical dysfunction.

The present study has important implications for the neurobiology of depression and anxiety. For example, the valence-arousal model is a neuropsychological model of emotion and psychopathology consisting of separate valence and arousal dimensions (Heller, 1993). In the original model, the valence dimension was associated with activity in frontal brain regions (with positive emotions implemented by the left frontal region and negative emotions implemented by the right frontal region) and the arousal dimension was associated with activity in the right parietal region. Depression and anxiety were posited to be associated with differing patterns of frontal and parietal brain activity, specifically decreased left frontal and right parietal activity in depression and increased right frontal and right parietal activity in anxiety. The valence-arousal model (Heller, 1993; Heller et al., 1997) has been well-supported regarding its prediction of decreased and increased right parietal activity in depression and anxiety, respectively. However, depression and anxiety are highly comorbid (Kessler et al., 2005) and it is less clear what pattern of right parietal activity to expect in those with both conditions.

The valence-arousal model appears to suggest that the opposing effects of depression and panic will cancel each other out, resulting in a “normal” level of right parietal activity for those with both conditions (Heller & Nitschke, 1998). However, the present study found that comorbid MDD/PD was associated with impaired performance on a primarily right parietal task (JLO) relative to MDD-only, PD-only, and control participants (who did not differ). These results are in contrast with the conjectured “normal” level of right parietal activity in comorbid depression and anxiety, but are consistent with prior parietal EEG asymmetry (Bruder et al., 1997; Kentgen et al., 2000; Mathersul et al., 2008; Metzger et al., 2004) and dichotic listening studies (Bruder et al., 1999; Pine et al., 2000), which found that comorbid depression and anxiety differed from depression only and controls (although see Bruder et al., 2004). Nelson and colleagues (2012) also found that comorbid depression and anxiety was associated with impaired performance on a primarily right frontal task (design fluency) relative to depression only and controls. Taken together, evidence suggests that the functional capabilities of the right hemisphere may be compromised in comorbid depression and anxiety, and that the presence of both conditions does not result in “normal” brain functioning.

The left parietal region has been largely ignored in theoretical and neuropsychological models of depression and anxiety (Shankman & Klein, 2003). Neuroimaging studies have reported left parietal dysfunction in depression (Müller et al., 2013) and PD (Meyer et al., 2000; Nordahl et al., 1990, 1998). The present study found impaired mathematical performance (a predominately left parietal function) in depression and anxiety, and it is possible that the underlying cerebral dysfunction contributes to other neuropsychological impairments. For example, the left parietal region has also been implicated in memory, language, and social cognition (Müller et al., 2013). Future research is needed to more definitely determine the role of left parietal dysfunction (and subsuming functions) in depression and anxiety.

Research on the cerebral lateralization of emotion has predominately focused on frontal asymmetry and the relative contribution of approach and withdrawal motivation in those with depression and/or anxiety (Davidson, 1992, 1998). The present study highlights the importance of parietal dysfunction in depression and anxiety and specifically the right hemisphere hypothesis of emotion that arose from the study of stroke patients (e.g., Borod et al., 1986). This hypothesis posits that the right hemisphere is specialized for processing emotions (Borod et al., 1998) and dysfunction in this area of the brain is related to affective psychopathology. However, research has been mixed as to whether strokes in the right parietal region are associated with post-stroke depression (Carson et al., 2000; Narushima et al., 2003). Future research is needed to better understand the role of parietal regions in the development and maintenance of depression and anxiety.

It is important to note that the present study was not designed to tease apart the sub-operations that may be disrupted in visuospatial processing and math abilities in depression and anxiety. There are several cognitive functions that may have been disrupted and contributed to the pattern of results. For example, poor JLO performance could be attributed to deficits in attentional capacity or visuospatial perception. Furthermore, poor WRAT-Arithmetic performance could have been the result of deficits in arithmetic facts, calculation procedures, or lexical processing (McCloskey, 1992). There also could have been impairment in particular working memory functions that are implemented by the parietal cortex, including inhibition, shifting, and updating (Collette et al., 2005). Future research is therefore needed to better understand the more specific cognitive functions that are impaired in these more global deficits.

The present study had several limitations that warrant consideration. First, and most importantly, WRAT-Arithmetic and JLO are not uniquely associated with left and right parietal cortices (Ng et al., 2000; Pinel & Dehaene, 2009) and it is possible that dysfunction in other brain regions may have contributed to the pattern of results. Second, the present study did not include effort testing, and it is possible that decreased performance was due to low motivation and not a lateralized deficit in neurocognitive abilities. However, we did not find diagnostic group differences on the number of items attempted on WRAT-Arithmetic. We also did not include a measure of premorbid intellectual capabilities, and it is possible that decreased performance was due to group differences in IQ. Third, the neuropsychological tasks were presented in a fixed order (JLO presented before WRAT-Arithmetic), and it is possible that order effects may have influenced the results. Finally, PD-only and comorbid MDD/PD participants were allowed to meet criteria for other anxiety disorders – thus, the findings for PD may have been due to anxiety in general, rather than PD specifically. However, given the large comorbidity among anxiety disorders, only requiring one anxiety disorder may have resulted in a less representative sample (Kessler et al., 2005).

In conclusion, the present study indicated visuospatial deficits in comorbid MDD/PD and mathematical impairment in MDD and/or PD. Importantly, the results were not accounted for by comorbid substance disorders, psychiatric medication use, or psychopathology severity. The present study highlights the importance of assessing parietally mediated cognitive functions in depression and anxiety, and future research is needed to better characterize the specific cognitive deficits that are common and unique to depression and anxiety disorders.