Neuropsychologia 143 (2020) 107486

Contents lists available at ScienceDirect

Neuropsychologia
journal homepage: http://www.elsevier.com/locate/neuropsychologia

Cortical surface area of the left frontal pole is associated with visuospatial
working memory capacity
George Zacharopoulos a, b, *, Torkel Klingberg b, 1, Roi Cohen Kadosh a, 1
a
Wellcome Centre for Integrative Neuroimaging, Department of Experimental Psychology, University of Oxford, UK
b
Department of Neuroscience, Karolinska Institute, Sweden

A R T I C L E I N F O A B S T R A C T

Keywords: Working memory, the ability to maintain and manipulate information over seconds, is central to cognition and it
Visuospatial working-memory capacity is impaired in many clinical populations. However, our knowledge of the structural properties associated with
Frontal pole individual variation in visuospatial working memory capacity is currently poor. Across two locations (Stockholm
Surface area
and Oxford), we examined how regional surface area and cortical thickness in frontal and parietal regions were
related to visuospatial working memory capacity. We found a negative association between visuospatial working
memory capacity and the surface area of the left frontal pole across both locations, and this finding was
consistently present in each of the two locations separately. Importantly, this association was specific to (i) the
surface area (not cortical thickness), (ii) the left side of the brain, (iii) and the visuospatial rather than the verbal
modality. This result reveals a novel and highly specific neurobiological association with visuospatial working
memory which could be further explored in studies with a wider range of psychological tests and in clinical
populations.

1. Introduction simulation and functional neuroimaging studies in humans. There is

still, however, less data on structural associations to vsWM capacity.
Working memory (henceforth WM) is our ability to maintain and Cortical morphology can be assessed with a volumetric representa­
manipulate information over seconds (Baddeley, 1986), and individual tion of the brain implemented by the technique of voxel-based
differences in WM capacity are large and seem to be highly stable over morphometry (Ashburner and Friston, 2000). Compared to
time (Kane and Engle, 2002). WM is a pivotal factor for cognition voxel-based morphometry (Hutton et al., 2009). Surface-based
(Conway et al., 2003; Süß et al., 2002). Individual variation in WM is a morphometry (Dale et al., 1999) separate the measurements of surface
crucial factor in a wide range of cognitive domains including reasoning, area and thickness, which is suggested to be more sensitive and specific
attention, mathematics and language comprehension (Alloway and measurements. Moreover, morphological measures such as cortical
Alloway, 2010; Conway et al., 2003). WM impairments have been thickness and surface area are among the parameters with the highest
documented in several clinical populations (Cools & D’Esposito, 2011) reliability in neuroimaging (interclass correlation>0.8) (Zuo et al.,
including Down’s syndrome, Williams syndrome, specific language 2019). Therefore, focusing on these structural parameters significantly
impairment, and attentional deficits (Gathercole and Alloway, 2006; optimizes measurement reliability which is often underappreciated in
Luck and Vogel, 2013). Consequently, there has been great interest in neuroscience research (Zuo et al., 2019).
uncovering the neurobiological underpinnings of WM capacity. The A study (Colom et al., 2013) of young-adults that investigated the
characterization of WM limitations can be described in terms of the brain morphological overlap between intelligence and cognitive factors
number of items an individual can maintain in mind successfully, such as identified cortical grey matter volume and cortical surface area in the
is measured by the K-value (Luck and Vogel, 1997), A current review middle frontal gyrus to be associated with both fluid intelligence and
identified a fronto-parietal network as key to visuospatial (henceforth WM, assessed using verbal and visuospatial subtests. Longitudinal
vs) WM capacity (Constantinidis and Klingberg, 2016). This review was developmental work (Tamnes et al., 2013) on children and adolescents
based on research from animal neurophysiology, neural network (8–19 years-old) revealed that improvement in verbal WM was related

* Corresponding author. Department of Experimental Psychology, University of Oxford, Oxford, OX2 6GG, United Kingdom.
E-mail address: [email protected] (G. Zacharopoulos).
1
Equal contribution.

https://doi.org/10.1016/j.neuropsychologia.2020.107486
Received 21 January 2020; Received in revised form 22 March 2020; Accepted 3 May 2020
Available online 11 May 2020
0028-3932/© 2020 Published by Elsevier Ltd.
G. Zacharopoulos et al. Neuropsychologia 143 (2020) 107486

to cortical volume reduction in bilateral prefrontal and posterior parietal were: (i) to identify the structural determinant of vsWM capacity, (ii) to
regions and in regions around the central sulci. Another developmental establish the structural, regional, and cognitive specificity of such a
study with children (from 6 to 16 years old) on the latent components of determinant.
WM (verbal, visuospatial and executive component) found no relation­
ship between cortical thickness and the verbal or vsWM components 2. Material and methods
(Bathelt et al., 2018) but found that the executive component of WM
could be predicted from an interaction between age and fractional 2.1. Participants
anisotropy within a corpus callosum and an occipitotemporal region. In
a structural investigation with healthy older adults, high verbal WM We recruited 39 participants (17 in Stockholm, 22 in Oxford) pre­
(n-back) performers exhibited more cortical surface area in the medial dominantly young adults and university students (Stockholm: mean age
orbital frontal gyrus, the inferior frontal gyrus and the superior frontal ¼ 25.5, standard deviation ¼ 3.6, 9 males, Oxford: mean age ¼ 26.05,
gyrus in the right hemisphere (Nissim et al., 2017). Another study with a standard deviation ¼ 6.5, 16 males). Participants were informed that the
large adult sample identified a negative association between cortical study investigated the behavioural and neural mechanisms of spatial
thickness within the middle frontal gyrus and visual WM performance memory. The completion of the structural acquisition lasted ~20min,
(Owens et al., 2018). and the completion of the WM testing lasted ~15min in Oxford and
Despite our understanding of the neurostructural properties of WM ~20min in Stockholm because of the additional behavioural measures
capacity, the prior work mainly focused primarily on visual and verbal (see below). Participants received monetary compensation for their
WM capacity and so very little knowledge exists on the specific brain participation. Informed written consent was obtained and the study was
regions and structural properties that track individual variation in approved by the Stockholm’s Ethics Committee and by the University of
vsWM capacity. Importantly, several findings suggest that there may be Oxford’s Medical Sciences Interdivisional Research Ethics Committee
distinct neural determinants between vsWM and other types of WM. (MS-IDREC-C2_2015_016), respectively.
First, several lesion studies revealed a regional dissociation between
vsWM and other types of WM and even between visual and spatial WM. 2.2. MRI data acquisition and pre-processing
Lesion in the right, but not left, parietal cortex induced impairments in
spatial WM (Koenigs et al., 2009), while verbal WM impairments were In Stockholm, the anatomical high-resolution T1-weighted volume
affected in patients with left lesions (Shallice and Warrington, 1970; scans were acquired using a BRAVO sequence (repetition time (TR) ¼
Vallar and Baddeley, 1984). Other studies showed that lesions to the 6.4040s; echo time (TE) ¼ 2.8080 ms, 180 slices; voxel size ¼ 1 � 1 � 1
temporal cortex affect visual but not spatial WM (Owen et al., 1996), mm) in a 3 T General Electrics MRI system. In Oxford, the anatomical
and the reverse pattern was documented in parietal-lesion patients high-resolution T1-weighted volume scans (1 mm3) were acquired with
(Pisella et al., 2004). Second, investigations of animal models indicated a 3 T Siemens MAGNETOM Prisma MRI System equipped with a 32
a dorsal/ventral dissociation based on spatial vs. object WM (Levy and channel receiver only head coil. Anatomical high-resolution T1-
Goldman-Rakic, 2000). Third, meta-analytic studies revealed that the weighted scans were acquired (MPRAGE sequence: TR ¼ 1900 ms; TE ¼
left prefrontal cortex (PFC) is engaged in verbal WM and the right PFC in 3.97 ms; 192 slices; voxel size ¼ 1 � 1 � 1 mm). Since the two locations
spatial WM (Nee et al., 2012; Owen et al., 2005; Wager and Smith, differed in their image acquisition parameters we controlled for location
2003). Fourth, other studies showed that different types of WM may when applicable as discussed below.
have different developmental trajectories and may be associated with The structural analysis was performed by utilising the recon-all
different cognitive abilities. For example, a dissociation between function within the FreeSurfer image analysis software v6.0.0, which
mastering auditory and visual WM tasks has led to the suggestion of an is documented and freely available for download online (surfer.nmr.
earlier functional maturity for the visual than for the auditory WM mgh.harvard.edu). The cortical parcellation of our regions of interest
system (Vuontela et al., 2003). Another developmental study (Swanson, was defined based on the Desikan-Killiany atlas. Our main structural
2017) indicated that performance on vsWM tasks as a function of age parameters of interest were cortical thickness (mm) and surface area
decreased at a faster rate than performance on verbal WM tasks. (mm2). Based on the previous studies mentioned in the introduction, we
Moreover, an investigation employing confirmatory factor analyses in a focused our analyses on frontoparietal regions which were the
study involving the measurement of verbal and spatial WM provided following: caudal middle frontal, inferior parietal, lateral orbitofrontal,
evidence of the need to distinguish verbal and spatial WM abilities (Hale medial orbitofrontal, pars opercularis, pars orbitalis, pars triangularis,
et al., 2011). Lastly, a study revealed that different types of WM were
related to different cognitive functions, in that vsWM was related to
mathematics while verbal WM was related to reading (Giofr� e et al.,
2018).
These accumulated findings raise the possibility that vsWM capacity
may stem from neurobiological mechanisms distinct from those of other
WM modalities. This possibility stresses the importance of investigating
the specific neurostructural underpinnings of vsWM capacity and its
corresponding limitations. Given the aforementioned previous studies,
an emerging open question is whether vsWM capacity resides in struc­
tural variation in key regions involved in WM in general (including
frontal and parietal). Therefore, the present study aimed to examine the
role of brain properties such as surface area and cortical thickness in
predicting vsWM capacity. To this end, we employed a two-location
study (Stockholm and Oxford) involved a scanning and a behavioural
session. During the scanning session, we obtained a structural image that
allowed the quantification of the structural properties of surface area
Fig. 1. Graphical representation of a single trial across the sequential trial
and cortical thickness. During the behavioural session, participants events: fixation, encoding, delay, probe and response phases (up to 2s). Par­
completed a vsWM task (Fig. 1) based on which we quantified the vsWM ticipants were asked to remember the location of 3, 4, 5 or 6 stimuli (red dots,
capacity. Lastly, in Stockholm participants completed another vsWM as displayed here). (For interpretation of the references to colour in this figure
task, the span board, and one verbal WM task, the digit span. Our aims legend, the reader is referred to the Web version of this article.)

2
G. Zacharopoulos et al. Neuropsychologia 143 (2020) 107486

rostral middle frontal, superior frontal, superior parietal, supramarginal,

frontal pole. In addition to the local measures, we also calculated the mean ​ K value ​ ðvsWMÞ ​ � ​ regional ​ surface ​ area ​ þ ​ total ​ surface ​ area ​
total global surface area and the mean cortical thickness for each þ ​ study ​ location (eq1.1)
individual.
mean ​ K value ​ ðvsWMÞ ​ � ​ regional ​ cortical ​ thickness ​
2.3. WM capacity measures þ ​ total ​ cortical ​ thickness ​ þ ​ study ​ location (eq1.2)
Out of these 24 regions, only the surface area of the left frontal pole
WM capacity for each participant was assessed in a single session.
exhibited a significant association with vsWM capacity (eq (1.1), Fig. 2B,
During the main vsWM task (Fig. 1) participants were presented with
B ¼ -.011, t ¼ 4.035, SE¼.003, PCORRECTED ¼ .0048, CI ¼ [-0.016,
three, four, five or six items followed by a delay and then the probe
-.006], uncorrected P-value ¼ .0001), even after controlling for age and
phase where they were required to make a yes/no button press
gender (eq (1.3), B ¼ -.009, t ¼ 3.457, Bias ¼ -.000231, SE¼.003,
depending on whether the encoding stimuli had appeared at the location
CI ¼ [-0.015–.004], uncorrected P-value ¼ .0023).
indicated by the probe stimulus. Participants completed 10 trials for
each of the four conditions (i.e., three, four, five or six items). The vsWM mean ​ K value ​ ðvsWMÞ ​ � ​ regional ​ cortical ​ thickness ​
capacity was measured using a standard formula (Cowan, 2001; Vogel þ ​ total ​ cortical ​ thickness ​ þ ​ study ​ location ​ þ ​ age ​ þ ​ gender
et al., 2005), K– – S*(H–F), where K is the vsWM capacity, S is the array
(eq1.3)
size, H is the observed hit rate and F is the false alarm rate. We calculated
the K for all four array sizes and average them into a single score rep­ Subsequently, we analysed the surface area of the left frontal pole in
resenting the vsWM capacity. each of the two study locations separately to assess whether this finding
In Stockholm, participants completed the forward versions of two was consistently present, which was indeed the case (Fig. 2C, Stock­
additional WM capacity tests to assess the specificity of our findings: (i) holm, B ¼ -.012, t ¼ -3.066, SE ¼ .004, P ¼ .006, CI ¼ [-0.018, -.002];
the span-board forward (Wechsler, 1981), which measures vsWM, and 2D, Oxford, B ¼ -.009, t ¼ 2.273, SE¼.004, P ¼ .034, CI ¼ [-0.018,
(ii) the digit span, which measures verbal WM. In the span-board, a -.001]. Of note, the main effect of the left frontal pole surface area was
sequence of blocks (e.g., 6,7,8) positioned on brick was indicated by the not significant (P ¼ .159) in predicting vsWM without total surface area
experimenter and the participant was asked to reproduce the sequence correction. For completeness, see Supplementary Material for the sta­
in the same order. The test stopped after two consecutive incorrect re­ tistical results of eq1 for all 48 brain regions.
sponses and the final score was calculated. In the digit span forward After establishing a reliable predictor of vsWM capacity, we exam­
(Wechsler, 2008), participants were asked to keep in memory a list of ined the specificity of this finding in terms of the (i) structural property
digits (e.g., 5, 8, 2) presented aurally at a rate of approximately one digit (surface area, cortical thickness), the (ii) side of the brain (left, right)
per second and were asked to immediately repeat the list in the same and the (iii) modality of WM (visuospatial, non-visuospatial). To
order. If they succeeded, a list one digit longer was presented. If they investigate the extent to which the association between vsWM capacity
failed, a second list of the same length was presented. If subjects were and left frontal pole is specific to the property of surface area as opposed
successful on the second list, a list one digit longer was given, as before. to the property of cortical thickness we added the left frontal pole
The test stopped after two consecutive incorrect responses and the final cortical thickness as well as the total cortical thickness (i.e., mean
score was calculated. cortical thickness across the whole brain) into the original model,
yielding eq (2). The addition of regional cortical thickness also requires
2.4. Statistical analyses the addition of total cortical thickness. Adding both the regional and
total structural properties (surface area and thickness) in the same
We employed multiple regression models in SPSS version 25 with the standard linear model allows us to examine the unique contribution of
WM capacity score as the dependent variable. The P-values and the 95% each on the vsWM.
confidence intervals (CI) presented in the results section were obtained mean ​ K value ​ ðvsWMÞ ​ � ​ regional ​ surface ​ area ​
from 10,000 samples bootstrapping methods (Ibm, 2017). The Boot­
þ ​ total ​ surface ​ area ​ þ ​ regional ​ thickness ​ þ ​ total ​ thickness ​
strapping method was used for deriving robust estimates of standard
errors and confidence intervals for the regression coefficient of the þ ​ study ​ location
neurostructural independent variables (see below). The independent (eq2)
variables were: (i) the regional cortical structure (surface area or cortical
The left surface area of the frontal pole was still the only regional
thickness within a region defined by the anatomical atlas); (ii) the global
significant predictor (B ¼ -.011, t ¼ 4.177, SE¼.002, P ¼ .0004,
cortical structure (global surface area or mean cortical thickness), and
CI ¼ [-0.015, -.007]). To examine if the association between vsWM ca­
(iii) the study location (Stockholm or Oxford). We included global
pacity and the frontal pole was specific to the left as opposed to the right
measures as covariates to correct for inter-individual differences in brain
hemisphere, we added the right frontal pole surface area into the orig­
size and to separate local and global signals.
inal model, yielding eq (3). Adding both the frontal lobe of both the right
We included the study location in the models as a dichotomous co­
and the left hemispheres in the same standard linear model allows us to
variate mainly because the two datasets were acquired using different
examine the unique contribution of each on the vsWM.
scanners with different image sequences and acquisition parameters.
Our main analyses involved combining the two datasets to improve mean​ K value​ ðvsWMÞ ​ � ​ left​ regional​ surface ​ area ​
power, as the sample size were small for each site separately and there þ ​ right ​ regional ​ surface ​ area​ þ ​ total​ surface ​ area ​ þ ​ study​ location
were gender ratio differences. To correct for multiple comparisons we (eq3)
utilised Bonferroni correction with 48 comparisons (24 brain measures
across the whole brain *2 structural properties), yielding any uncor­ Again, the surface area of the left frontal pole was the only regional
rected P-value below .001042 to be significant (i.e., PCORRECTED<.05). significant predictor (B ¼ -.010, t ¼ 3.952, SE¼.002, P ¼ .0002,
CI ¼ [-0.015, -.006]).
3. Results Our last aim was to establish the cognitive specificity of the main
findings. The following analyses were conducted only in the Stockholm
We first examined the association between vsWM capacity and the dataset as the additional WM tasks were collected only for this location.
surface area (eq (1.1)) and cortical thickness (eq (1.2)) of 12 frontal and For this reason, the following equations do not involve the predictor
parietal regions (per hemisphere), using eq1. study location. To confirm that the surface area of the left frontal pole is

3
G. Zacharopoulos et al. Neuropsychologia 143 (2020) 107486

Fig. 2. Scatterplots depicting the associations be­

tween individual variation in the left frontal pole
surface area. (A) Parcelleated left hemisphere where
the frontal pole is displayed in the most anterior part
with a dark purple colour. The y-axis displays the
surface area of the left frontal pole (here the unstan­
dardized residuals when correcting for total surface
area and study side) and the x-axis represents the
vsWM capacity measured as the mean K-value (B)
across locations, (C) in Stockholm, and (D) in Oxford
locations separately. (For interpretation of the refer­
ences to colour in this figure legend, the reader is
referred to the Web version of this article.)

associated with vsWM, we assessed whether the surface area of the left neurostructural underpinnings of WM capacity but our knowledge of the
frontal pole can track individual variation of another vsWM test, namely specific neurostructural properties contributing to vsWM capacity, in
the span board, using eq (4), which was indeed the case (B ¼ -.026, particular, is very scarce. Here, we identified that individuals with
t ¼ 3,174, SE¼.001, P ¼ .006, CI ¼ [-0.038, -.006]), even after addi­ relatively smaller surface area (i.e., controlling for the total surface area)
tionally controlling for age and gender (B ¼ -.023, t ¼ 2.714, within the left frontal pole exhibited higher vsWM capacity, and this
Bias¼.002, SE¼.009, CI ¼ [-0.038–.001], uncorrected P-value ¼ .0027). finding was consistently present in each the two research locations
separately. Our findings seem to be in line with prior microstructural
Span ​ board ​ ðvsWMÞ ​ � ​ regional ​ surface ​ area ​ þ ​ total ​ surface ​ area
work, which also found the involvement of frontal regions in tracking
(eq4) vsWM. For example, a study showed that structural integrity in fronto­
To investigate if the surface area of the left frontal pole is associated parietal areas is correlated with vsWM capacity (Klingberg, 2006), and
with verbal WM we employed the digit span forward test using eq (5) fractional anisotropy values within frontoparietal regions were posi­
but the main effect of the left frontal pole did not reach significance tively correlated with vsWM performance and WM-related functional
(B ¼ -.025, t ¼ 2.073, SE¼.015, P ¼ .07, CI ¼ [-0.053, 0.008]). activity (Olesen et al., 2003).
However, the exact neurobiological nature of the role of the left
Digit ​ span ​ ðverbal ​ WMÞ ​ � ​ regional ​ surface ​ area ​ þ ​ total ​ surface ​ area frontal pole’s surface area on vsWM capacity is currently unknown. The
(eq5) cortical surface area is thought to reflect the structural integrity of grey
Lastly, using eq (6) the significance of the left frontal pole in pre­ matter (Fischl and Dale, 2000; Lemaitre et al., 2012; Salat et al., 2004).
dicting vsWM was present even after controlling for the verbal WM In particular, neurons within the cerebral cortex are organized into
(B ¼ -.015, t ¼ 3.221, SE¼.005, P ¼ .01, CI ¼ [-0.023, -.002]). columns that are perpendicular to the surface of the brain. The surface
area is likely associated with the number of radial columns perpendic­
Mean ​ K value ​ ðvsWMÞ ​ � ​ regional ​ surface ​ area ​ þ ​ total ​ surface ​ area ​ ular to the pial surface whereas cortical thickness is defined by the
þ ​ digit ​ span ​ ðverbal ​ WMÞ (eq6) horizontal layers in the cortical columns (Rakic, 2009). Indeed, the
radial unit hypotheses postulate that the size of the cortical surface area
Taken together, these results show that the surface area of the left
is determined by the number of cortical columns (Rakic, 1995, 2007;
frontal pole is specific to the visuospatial modality of WM.
Rakic and Swaab, 1988), and several proteins, such as β-catenin and
caspase 3/9, have been shown to be significant contributors of surface
4. Conclusions
area. Indeed, increases in surface area have been documented with both
the stabilized β-catenin transgene (Chenn and Walsh, 2003) and with
In the present study, we examined the association between structural
caspase 3/9 mutations which lead to a larger number of radial columns
properties of the brain and vsWM capacity. This was done by combining
and cause decreased apoptosis of progenitor cells and radial glial cells
brain imaging, allowing the quantification of several structural prop­
respectively (Haydar et al., 1999). It is currently difficult based on the
erties (surface area, cortical thickness) with a computerised task
present study alone to make inferences as to which of these proteins may
assessing vsWM capacity. We reported three main findings. First, we
affect the surface area of the left frontal pole resulting in the formation of
identified the role of the surface area within the left frontal pole in
vsWM capacity. It was also suggested that surface area measures may
tracking vsWM capacity, which was independent of gender and age.
even reflect the underlying white matter fibres with tension or shrinkage
Second, we found that the structural property of the left surface area was
of these fibers leading to deeper sulci and extended surface area mea­
significantly associated with vsWM capacity while the structural prop­
sures (Van Essen, 1997). This may suggest that shrinkage of the white
erty of cortical thickness was not. Third, the surface area of the left
fibres that reach the left frontal pole may be more shrunk in the in­
frontal pole tracked individual variation in vsWM but not verbal WM,
dividuals with lower vsWM capacity scores. Despite these speculations,
indicating modal specificity.
future research incorporating a multi-modal approach combining
WM capacity is a central factor of cognitive functioning and,
structural and functional connectivity indices can detail the neural
consequently, WM capacity limitations impair cognitive functioning.
network through which the surface area of the left frontal pole may
Therefore, a central aim of neuroscience research is the characterization
impact the brain network level forming vsWM capacity. A limitation of
of the neurobiological underpinnings of such limitations. As mentioned
the present study is that the frontal pole was treated as a unitary region.
in the introduction, several previous studies explored the

4
G. Zacharopoulos et al. Neuropsychologia 143 (2020) 107486

However, the frontal pole is subdivided into lateral (Fp1) and medial experiments; G.Z. analysed the data, G.Z., T.K., R.C.K.wrote the paper.
(Fp2) cytoarchitectonic subregions, which were shown to coactivate
with a different set of regions as assessed using fMRI (Bludau et al.,
2014). Regions involved in top-down control such as the dorsolateral Declaration of competing interest
prefrontal cortex and anterior cingulate cortex co-activate with the
lateral frontal pole while medial frontal pole coactivates with temporal The authors declare no competing financial interests.
pole, posterior cingulate and posterior superior temporal sulcus (Gilbert
et al., 2010). Taken these findings together, the frontal pole is thought to Acknowledgements
be an attention control system responsible to select between attending to
internal (lateral frontal pole) or attending to external (medial frontal We are grateful to all the participants who took part in our study, and
pole) information collectively termed the gateway hypothesis (Orr et al., Paul Haggar for helping in editing the manuscript. The Wellcome Centre
2015). Indeed, succeeding in the present vsWM task requires switching for Integrative Neuroimaging is supported by core funding from the
between attending to external (stimulus encoding phase) and attending Wellcome Trust (203139/Z/16/Z). The Oxford location work was sup­
to internal (maintenance phase), and back to external information ported by the European Research Council (Learning & Achievement
(response probe). Indeed prior resting-state functional connectivity 338065).
study revealed that the lateral frontal pole is connected to the executive
control network, associated with directed attention and WM (Moayedi Appendix A. Supplementary data
et al., 2014). There is also evidence suggesting that the lateral frontal
pole is uniquely human as connectivity comparison between humans vs. Supplementary data to this article can be found online at https://doi.
macaques using structural and functional neuroimaging methods org/10.1016/j.neuropsychologia.2020.107486.
revealed that the ventrolateral frontal pole is specific to humans as it
could not match any prefrontal cortex macaque region (Neubert et al.,
References
2014). Based on these findings we speculate that the surface area of the
lateral frontal pole than the total frontal pole area should be more Alloway, T.P., Alloway, R.G., 2010. Investigating the predictive roles of working memory
relevant to vsWM capacity. and IQ in academic attainment. J. Exp. Child Psychol. 106 (1), 20–29.
Ashburner, J., Friston, K.J., 2000. Voxel-based morphometry—the methods. Neuroimage
Apart from the medial-lateral distinction recent evidence using
11 (6), 805–821.
anatomical and functional connectivity methodologies suggest the ex­ Baddeley, A., 1986. Working Memory. Oxford University Press, New York.
istence of ventral-dorsal gradient within the frontal pole (Orr et al., Bathelt, J., Gathercole, S.E., Johnson, A., Astle, D.E., 2018. Differences in brain
2015). According to this view, dorsal regions are connected to process morphology and working memory capacity across childhood. Dev. Sci. 21 (3),
e12579.
goals and action plan prefrontal regions, medial regions are connected to Bludau, S., Eickhoff, S.B., Mohlberg, H., Caspers, S., Laird, A.R., Fox, P.T., Amunts, K.,
regions monitoring action outcomes and motivating behaviour and 2014. Cytoarchitecture, probability maps and functions of the human frontal pole.
ventral regions connect to regions processing information about stimuli, Neuroimage 93, 260–275.
Chenn, A., Walsh, C.A., 2003. Increased neuronal production, enlarged forebrains and
values and emotions. However, it is still unclear which of this set of cytoarchitectural distortions in β-catenin overexpressing transgenic mice. Cerebr.
frontal pole connections are most relevant to vsWM. Future structural Cortex 13 (6), 599–606.
and connectivity work can examine whether the structural parameters Colom, R., Burgaleta, M., Rom� an, F.J., Karama, S., Alvarez-Linera,
� J., Abad, F.J.,
Haier, R.J., 2013. Neuroanatomic overlap between intelligence and cognitive
or connectivity patterns of specific frontal pole’s subregions is more factors: morphometry methods provide support for the key role of the frontal lobes.
critical in tracking vsWM capacity. In particular, such studies can reveal Neuroimage 72, 143–152.
the impact of the left frontal pole surface area at the mechanistic Constantinidis, C., Klingberg, T., 2016. The neuroscience of working memory capacity
and training. Nat. Rev. Neurosci. 17 (7), 438.
network level, a piece of missing information that we could not address
Conway, A.R., Kane, M.J., Engle, R.W., 2003. Working memory capacity and its relation
with our neuroimaging protocol. For instance, the effect of frontal pole’s to general intelligence. Trends Cognit. Sci. 7 (12), 547–552.
surface area on vsWM capacity may be mediated by the connectivity Cools, R., D’Esposito, M., 2011. Inverted-U–shaped dopamine actions on human working
memory and cognitive control. Biol. Psychiatr. 69 (12), e113–e125.
between left lateral frontal pole and seeds within the executive control
Cowan, N., 2001. The magical number 4 in short-term memory: a reconsideration of
network. This is a likely case as previous studies established these con­ mental storage capacity. Behav. Brain Sci. 24 (1), 87–114.
nections to frontal pole, and these regions are known to underpin WM Dale, A.M., Fischl, B., Sereno, M.I., 1999. Cortical surface-based analysis: I. Segmentation
(Moayedi et al., 2014). and surface reconstruction. Neuroimage 9 (2), 179–194.
Fischl, B., Dale, A.M., 2000. Measuring the thickness of the human cerebral cortex from
Another limitation of the current investigation is the fact that the magnetic resonance images. Proc. Natl. Acad. Sci. Unit. States Am. 97 (20),
main finding is of a correlation in nature. Although several specific 11050–11055.
proteins discussed above were shown to experimentally alter the surface Gathercole, S.E., Alloway, T.P., 2006. Practitioner review: short-term and working
memory impairments in neurodevelopmental disorders: diagnosis and remedial
area, measuring vsWM capacity in mice has its limitations and may not support. JCPP (J. Child Psychol. Psychiatry) 47 (1), 4–15.
be applicable. A more viable method to probe causality may involve Gilbert, S.J., Gonen-Yaacovi, G., Benoit, R.G., Volle, E., Burgess, P.W., 2010. Distinct
cognitive training, where individuals are trained in the vsWM and verbal functional connectivity associated with lateral versus medial rostral prefrontal
cortex: a meta-analysis. Neuroimage 53 (4), 1359–1367.
WM task described here and assessed as to whether this training inter­ Giofr�
e, D., Donolato, E., Mammarella, I.C., 2018. The differential role of verbal and
vention induced training alterations, specifically in the surface area of visuospatial working memory in mathematics and reading. Trends in neuroscience
the left frontal pole and in response to the visuospatial rather than the and education 12 (1), 1–6.
Hale, S., Rose, N.S., Myerson, J., Strube, M.J., Sommers, M., Tye-Murray, N., Spehar, B.,
verbal WM training.
2011. The structure of working memory abilities across the adult life span. Psychol.
In summary, by investigating cortical morphometric properties such Aging 26 (1), 92.
as surface area and thickness the present research allowed the identifi­ Haydar, T.F., Kuan, C.-Y., Flavell, R.A., Rakic, P., 1999. The role of cell death in
regulating the size and shape of the mammalian forebrain. Cerebr. Cortex 9 (6),
cation of an objective neurostructural marker for vsWM capacity. Given
621–626.
the central role of WM capacity on cognitive processes, these findings Hutton, C., Draganski, B., Ashburner, J., Weiskopf, N., 2009. A comparison between
pave the way for further studies into the causal role of the neural dif­ voxel-based cortical thickness and voxel-based morphometry in normal aging.
ferences influencing core components of cognition which are impaired Neuroimage 48 (2), 371–380.
Ibm, C., 2017. IBM SPSS Statistics for Windows, Version Q3 25.0. IBM Corp, Armonk,
in different clinical populations. NY.
Kane, M.J., Engle, R.W., 2002. The role of prefrontal cortex in working-memory capacity,
Contributions executive attention, and general fluid intelligence: an individual-differences
perspective. Psychon. Bull. Rev. 9 (4), 637–671.
Klingberg, T., 2006. Development of a superior frontal–intraparietal network for visuo-
G.Z., T.K., R.C.K. designed the experiments; G.Z. performed the spatial working memory. Neuropsychologia 44 (11), 2171–2177.

5
G. Zacharopoulos et al. Neuropsychologia 143 (2020) 107486

Koenigs, M., Barbey, A.K., Postle, B.R., Grafman, J., 2009. Superior parietal cortex is Pisella, L., Berberovic, N., Mattingley, J.B., 2004. Impaired working memory for location
critical for the manipulation of information in working memory. J. Neurosci. 29 but not for colour or shape in visual neglect: a comparison of parietal and non-
(47), 14980–14986. parietal lesions. Cortex 40 (2), 379–390.
Lemaitre, H., Goldman, A.L., Sambataro, F., Verchinski, B.A., Meyer-Lindenberg, A., Rakic, P., 1995. A small step for the cell, a giant leap for mankind: a hypothesis of
Weinberger, D.R., Mattay, V.S., 2012. Normal age-related brain morphometric neocortical expansion during evolution. Trends Neurosci. 18 (9), 383–388.
changes: nonuniformity across cortical thickness, surface area and gray matter Rakic, P., 2007. The radial edifice of cortical architecture: from neuronal silhouettes to
volume? Neurobiol. Aging 33 (3), 617 e611-617. e619. genetic engineering. Brain Res. Rev. 55 (2), 204–219.
Levy, R., Goldman-Rakic, P.S., 2000. Segregation of working memory functions within Rakic, P., 2009. Evolution of the neocortex: a perspective from developmental biology.
the dorsolateral prefrontal cortex. In: Executive Control and the Frontal Lobe: Nat. Rev. Neurosci. 10 (10), 724.
Current Issues. Springer, pp. 23–32. Rakic, P., Swaab, D., 1988. Defects of neuronal migration and the pathogenesis of
Luck, S.J., Vogel, E.K., 1997. The capacity of visual working memory for features and cortical malformations. In: Progress in Brain Research, vol. 73. Elsevier, pp. 15–37.
conjunctions. Nature 390 (6657), 279. Salat, D.H., Buckner, R.L., Snyder, A.Z., Greve, D.N., Desikan, R.S., Busa, E., Fischl, B.,
Luck, S.J., Vogel, E.K., 2013. Visual working memory capacity: from psychophysics and 2004. Thinning of the cerebral cortex in aging. Cerebr. Cortex 14 (7), 721–730.
neurobiology to individual differences. Trends Cognit. Sci. 17 (8), 391–400. Shallice, T., Warrington, E.K., 1970. Independent functioning of verbal memory stores: a
Moayedi, M., Salomons, T.V., Dunlop, K.A., Downar, J., Davis, K.D., 2014. Connectivity- neuropsychological study. Q. J. Exp. Psychol. 22 (2), 261–273.
based parcellation of the human frontal polar cortex. Brain Struct. Funct. 1–14. Süß, H.-M., Oberauer, K., Wittmann, W.W., Wilhelm, O., Schulze, R., 2002. Working-
https://doi.org/10.1007/s00429-014-0809-6. memory capacity explains reasoning ability—and a little bit more. Intelligence 30
Nee, D.E., Brown, J.W., Askren, M.K., Berman, M.G., Demiralp, E., Krawitz, A., (3), 261–288.
Jonides, J., 2012. A meta-analysis of executive components of working memory. Swanson, H.L., 2017. Verbal and visual-spatial working memory: what develops over a
Cerebr. Cortex 23 (2), 264–282. life span? Dev. Psychol. 53 (5), 971.
Neubert, F.-X., Mars, R.B., Thomas, A.G., Sallet, J., Rushworth, M.F., 2014. Comparison Tamnes, C.K., Walhovd, K.B., Grydeland, H., Holland, D., Østby, Y., Dale, A.M., Fjell, A.
of human ventral frontal cortex areas for cognitive control and language with areas M., 2013. Longitudinal working memory development is related to structural
in monkey frontal cortex. Neuron 81 (3), 700–713. maturation of frontal and parietal cortices. J. Cognit. Neurosci. 25 (10), 1611–1623.
Nissim, N.R., O’Shea, A.M., Bryant, V., Porges, E.C., Cohen, R., Woods, A.J., 2017. Vallar, G., Baddeley, A.D., 1984. Fractionation of working memory: neuropsychological
Frontal structural neural correlates of working memory performance in older adults. evidence for a phonological short-term store. J. Verb. Learn. Verb. Behav. 23 (2),
Front. Aging Neurosci. 8, 328. 151–161.
Olesen, P.J., Nagy, Z., Westerberg, H., Klingberg, T., 2003. Combined analysis of DTI and Van Essen, D.C., 1997. A tension-based theory of morphogenesis and compact wiring in
fMRI data reveals a joint maturation of white and grey matter in a fronto-parietal the central nervous system. Nature 385 (6614), 313.
network. Cognit. Brain Res. 18 (1), 48–57. Vogel, E.K., McCollough, A.W., Machizawa, M.G., 2005. Neural measures reveal
Orr, J.M., Smolker, H.R., Banich, M.T., 2015. Organization of the human frontal pole individual differences in controlling access to working memory. Nature 438 (7067),
revealed by large-scale DTI-based connectivity: implications for control of behavior. 500.
PloS One 10 (5). Vuontela, V., Steenari, M.-R., Carlson, S., Koivisto, J., Fj€
allberg, M., Aronen, E.T., 2003.
Owen, A.M., McMillan, K.M., Laird, A.R., Bullmore, E., 2005. N-back working memory Audiospatial and visuospatial working memory in 6–13 year old school children.
paradigm: a meta-analysis of normative functional neuroimaging studies. Hum. Learn. Mem. 10 (1), 74–81.
Brain Mapp. 25 (1), 46–59. Wager, T.D., Smith, E.E., 2003. Neuroimaging studies of working memory. Cognit. Affect
Owen, A.M., Morris, R.G., Sahakian, B.J., Polkey, C.E., Robbins, T.W., 1996. Double Behav. Neurosci. 3 (4), 255–274.
dissociations of memory and executive functions in working memory tasks following Wechsler, D., 1981. Wechsler Adult Intelligence Scale: WAIS-R Manual. Harcourt Brace
frontal lobe excisions, temporal lobe excisions or amygdalo-hippocampectomy in Jovanovich [for] The Psychological Corporation.
man. Brain 119 (5), 1597–1615. Wechsler, D., 2008. Wechsler Adult Intelligence Scale–Fourth Edition (WAIS–IV), vol.
Owens, M.M., Duda, B., Sweet, L.H., MacKillop, J., 2018. Distinct functional and 22. NCS Pearson, San Antonio, TX, p. 498.
structural neural underpinnings of working memory. Neuroimage 174, 463–471. Zuo, X.-N., Xu, T., Milham, M.P., 2019. Harnessing reliability for neuroscience research.
Nature human behaviour 3 (8), 768–771.