(1) Department of Anatomy, University of Athens School of Medicine, Athens, Greece
* Corresponding author Email:
The main purpose of our study was to measure the cortical thickness of the cortical connections of the human nucleus accumbens in order to explore potential morphometric correlations. Furthermore, we tested the hypothesis of a morphometric correlation between the nucleus accumbens and the cingulate gyrus.
Materials and methods
The material consisted of 41 cerebral hemispheres (25 left and 16 right) from 25 normal human brains. They were obtained from 22 males who were 50–90 years old, and from three females who were 67–94 years old. We measured the thickness of four cortical areas connected to the nucleus accumbens: the cingulate, entorhinal, orbitofrontal and piriform cortices, as well as the height of the subgenual part of the cingulate gyrus.
We found a very statistically significant correlation between the orbitofrontal and entorhinal cortices, significant correlation between the cingulate and orbitofrontal cortices, significant correlation between the piriform and orbitofrontal cortices and significant correlation between the piriform and entorhinal cortices.
Our study indicated that the cingulate cortex is probably the thickest cortical area connected to the nucleus accumbens. It also suggested a potentially more significant relation between the orbitofrontal cortex and the limbic system than what is currently believed. Furthermore, we provided evidence that the size of the nucleus accumbens is neither correlated with the thickness of its cortical connections nor with the size of the cingulate gyrus.
The human nucleus accumbens (NA), which belongs to the basal ganglia of the brain, is the main part of the ventral striatum. It is a round-shaped, dorsally flattened structure, symmetrically placed anterior to the anterior commissure (AC) and lies parallel to the midline. It covers a large area of the basal forebrain (Figure 1) and is the region of continuity between the putamen and the head of the caudate nucleus[1,2,3,4,5]. The NA is generally accepted to be located underneath the anterior limb of the internal capsule, laterally to the vertical part of the Broca’s diagonal band and medially to the claustrum and piriform cortex. It extends dorsolaterally into the putamen and dorsomedially into the caudate nucleus (Figure 2) without a sharp demarcation. The NA is chemically divided into two parts: a shell, laterally and a core, medially. The first is more related to the limbic system and the second to the extrapyramidal motor system. The NA, having dopamine as a major transmitter, is a crucial centre for the experience of reward and pleasure and is also considered to be a psychosurgery target[2,7]. It is also a part of the forebrain circuitry involved in the regulation of behavioural activation and effort-related decision-making.
Formalin-fixated human brain from a middle-aged male, left hemisphere, sagittal section, 7 mm lateral to the midline (coronal section 2 mm anterior to the AC) revealing the NA location. 1: NA, 2: internal capsule (anterior limb), 3: caudate nucleus (head), 4: lateral ventricle (frontal horn), 5: fibres of the corpus callosum (genu).
Formalin-fixated human brain from a middle-aged male, right hemisphere, transverse section at the AC-PC plane (coronal section 2 mm anterior to the AC) revealing the NA location. 1: NA, 2: diagonal band of Broca, 3: caudate nucleus (head), 4: internal capsule (anterior limb), 5: putamen, 6: external capsule.
Connections of the NA
The ventromedial striatum (containing the ventromedial caudate nucleus, NA core and NA shell) receives inputs through the anterior internal capsule (Figure 3) from the medial orbitofrontal cortex and the anterior and subgenual cingulate cortex (Brodmann areas lateral 10, 11, 12, 13, 24, 25 and 32). The NA receives mainly glutamatergic projections from the anterior cingulate cortex, frontal cortex, prefrontal cortex[8,10], amygdala[2,8,10] (central and medial amygdaloid nuclei), hippocampus[8,10] and thalamus, and a strong dopaminergic projection from the mesencephalon, i.e. ventral tegmental area (VTA)[2,8,10] and substantia nigra (SN).
Formalin-fixated human brain from a middle-aged male, left hemisphere, coronal section 2 mm anterior to the AC. 1: NA, 2: pinhead representing an electrode’s target point for NA deep brain stimulation, 3: caudate nucleus (head), 4: putamen, 5: internal capsule (anterior limb), 6: external capsule, 7: claustrum, 8: extreme capsule, 9: AC–PC plane, 10: corpus callosum, 11: septum pellucidum, 12: lateral ventricle (frontal horn).
The NA shell receives inputs from the prefrontal cortex (infralimbic, ventral agranular insular and ventral prelimbic cortices)[10,11], dorsal peduncular cortex, piriform cortex, medial and lateral entorhinal cortex, orbital cortex, amygdala (caudal basolateral amygdaloid and rostral basal amygdaloid complexes), hippocampus (ventral subiculum/cornu Ammonis field 1), thalamus (anterior and posterior paraventricular nucleus) and mesencephalon (medial VTA, lateral VTA/SN compact part/retrorubral nucleus).
The NA core receives inputs from the prefrontal cortex (dorsal agranular insular and dorsal prelimbic cortices)[10,11], perirhinal cortex, anterior cingulate cortex, premotor and supplemental motor cortices, amygdala (basal amygdaloid complex), thalamus (intermediodorsal nucleus/central medial nucleus) and mesencephalon (medial SN compact part).
The main efferents of the NA innervate the pallidum, striatum, mediodorsal thalamus, prefrontal and cingulate cortices and mesolimbic dopaminergic areas. The major efferent projection from the NA terminates in the ventral pallidum and is principally gamma-aminobutyric acid-ergic[8,10]. The ventral pallidum, in turn, projects strongly to the SN compact part (mediolateral part), as well as to the limbic part of the subthalamic nucleus (STN), and to the latter’s extensions into the local hypothalamus. It also projects to mediodorsal thalamus and various brainstem motor areas. It has been hypothesised that the ventral pallidum acts as a relay station and as an integrator of information related to diverse limbic and striatal inputs. In addition, the NA provides a recurrent projection to the VTA and SN.
The NA shell projects to the lateral hypothalamus, extended amygdala, ventrolateral and ventromedial (subcommissural) ventral pallidum[9,10,11] and dorsal SN reticular part/VTA. The NA core projects to the dorsolateral ventral pallidum[10,11] (internal part of the ventromedial globus pallidus), STN and dorsal SN reticular part/VTA. Both pathways terminate in the thalamus, which receives and sends fibres to the medial orbitofrontal cortex/anterior and subgenual cingulate cortex. An important difference between the NA core and the NA shell is the efferent projection from the NA shell to the lateral hypothalamus and the extended amygdala, which does not exist in the NA core.
The limbic cortico-striato-pallido-thalamo-cortical loop
The concept of the ventral striato-pallidal system provided the first indication for the existence of parallel cortico-striato-pallido-thalamo-cortical circuits, which in turn led to the theory of segregated cortical-subcortical re-entrant circuits as a conceptual framework for the study of psychiatric disorders. The NA, an integral and important part of the limbic and prefrontal cortico-striato-pallido-thalamic circuits, seems to function as a limbic-motor interface and is involved in several cognitive, emotional and psychomotor functions, which have been found to be altered in some psychopathological conditions.
The existence of a robust and direct striato-pallido-thalamo-cortico-striatal pathway from the shell to the core, suggests that significant co-ordination of the levels of activity in the two sub-territories should exist. Furthermore, the intimate connectional relations of the shell and core with the dorsal pre-limbic and dorsal agranular insular cortices are consistent with the role of the NA in the regulation of cognitive aspects of adaptive responding, possibly involving the gating of response initiation.
Purpose of the study
The main purpose of our study was to measure the cortical thickness of the NA cortical connections in order to explore potential morphometric correlations between the NA and each of the studied cortical areas, as well as among the individual cortical areas. Furthermore, we tested the hypothesis of a morphometric correlation between the NA and cingulate gyrus (CG) (subgenual part). The importance of the cingulate cortex as a NA connection is described in the discussion section of this article.
This work conforms to the values laid down in the Declaration of Helsinki (1964). The protocol of this study has been approved by the relevant ethical committee related to our institution in which it was performed.
The material consisted of 41 cerebral hemispheres (25 left and 16 right) from 25 normal human brains that we had in our department. They were obtained from 22 males, 50–90 years old and 3 females, 67–94 years old, who were cadaver donors for students’ education. These brains had been fixated in formalin solution.
We measured the thickness of four cortical areas connected to the NA: the cingulate, entorhinal, orbitofrontal and piriform cortices. Specifically, we chose the thickness of the cingulate cortex anterior to the genu of the corpus callosum (C), the orbitofrontal cortex thickness at the intercommissural (anterior commissure– posterior commissure, AC–PC) plane, 10 mm laterally off the midline (O), the piriform cortex thickness posterior to the olfactory tract (P) and the entorhinal cortex thickness at the midpoint of the hippocampal uncus (E). Methodologically, we used a transparent plastic tube of about 25 mm length and 2 mm diameter, inserted perpendicularly to the cerebral surface, to remove a cylindrical piece of brain tissue from each selected area. The transparency of the tube allowed us to identify the cortex-white matter limit and hence allowed us to measure the cortical thickness (Figure 4). A scalpel was occasionally used to help the removal of the brain tissue piece (Figure 4).
Instruments used for taking a cortical sample from brain regions. 1: tube, 2: cerebral cortex, 3: subcortical white matter, 4: scalpel, 5: scalpel blade.
We also measured the height of the subgenual part of the CG (CGH), 12 mm rostral to the anterior border of the AC, in order to test the hypothesis of an existing morphometric correlation between the NA and the CGH. We also tested the hypothesis of such a correlation between the NA and the C, O, P and E. As an index of the NA size, we used its maximum coronal dimension (diameter), 2 mm anterior to the AC (Dmax). Our methodology to measure Dmax has been previously published.
Table 1 presents our measurements. The C varied from 0.8 mm to 5 mm and its mean value (MV) ± standard deviation (SD) was C = 2.41 ± 1.07 mm (n = 26). For right hemispheres, we found C = 2.53 ± 1.61 mm (n = 9) and for left, C = 2.35 ± 1.29 mm (n = 17). The O varied from 1.1 mm to 5.0 mm and its MV ± SD was O = 2.16 ± 0.8 mm (n = 36). For right hemispheres, we found O = 2.27 ± 1.23 mm (n = 14) and for left, O = 2.09 ± 1.01 mm (n = 22). The P varied from 1.0 mm to 4.0 mm and its MV ± SD was P = 2.19 ± 0.74 mm (n = 31). For right hemispheres, we found P = 2.06 ± 1.25 mm (n = 12) and for left, P = 2.27 ± 1.24 mm (n = 19). The E varied from 1.0 mm to 5.0 mm and its MV ± SD was E = 2.11 ± 0.87 mm (n = 37). For right hemispheres, we found E = 2.13 ± 1.33 mm (n = 14) and for left, E = 2.10 ± 1.19 mm (n = 23). We found no statistically significant difference in the MVs of the C, O, P and E between the right and left hemispheres.
Measurements of the cingulate gyrus height and thickness of the cingulate, orbitofrontal, piriform and entorhinal cortices.
The CGH varied from 5 mm to 10 mm and its MV ± SD was CGH = 7.21 ± 1.26 (n = 29). For right hemispheres, we found CG = 7.00 ± 1.33 (n = 10) and for left, CGH = 7.32 ± 1.25 (n = 19). We found no significant differences in the CGH between the right and left hemispheres.
The statistical analysis of our measurements revealed the following correlations (beginning from the most powerful):
• Statistically very significant correlation between the O and E (r = 0.609, df = 32, p<0.001).
• Statistically significant correlation between the C and O (r = 0.540, df = 24, p<0.01).
• Statistically significant correlation between the P and O (r = 0.489, df = 29, p<0.01).
• Statistically significant correlation between the P and E (r = 0.425, df = 27, p<0.05).
We also found no statistical correlation between: Dmax and CGH (r = 0.277, df = 26, p > 0.1), Dmax and C (r = 0.084, df = 22, p > 0.1), Dmax and O (r = 0.240, df = 25, p > 0.1), Dmax and P (r = 0.007, df = 20, p > 0.1), and Dmax and E (r = 0.106, df = 25, p > 0.1).
To the best of our knowledge, our anatomical method for the measurement of cortical thickness has not been previously described. Our choice of these specific cortical points (C, O, P and E) was based on easy-to-locate areas of the specific brain surface areas studied, hence being easily reproducible. Except for the cingulate cortical thickness C, we also used the CGH as an index of the CG size. The cingulate cortex is a quite important connection of the NA because it is one of the very few brain areas connected to this nucleus with both afferent and efferent fibres[2,8]. We chose this particular part (CGH) of the CG because of its clearer limits, which provided easier and more reliable measurements, and because the CGH is the nearest to the NA part of this gyrus.
Considering the C, O, P, E and CGH, we would like to emphasise the absence of a significant morphometric correlation with the NA and also the absence of a significant interhemispheric difference. Regarding the mean thickness results, we observed that C>P>O>E (Table 1). Consequently, the cingulate cortex is probably the thickest cortical area connected to the NA (at least of those studied). The very significant correlation between the orbitofrontal and entorhinal cortices is quite an interesting finding, difficult to be explained. Together with the significant correlation between the cingulate and orbitofrontal cortices, it might suggest a more significant relation of the orbitofrontal cortex to the limbic lobe (where cingulate and entorhinal cortices belong) than what is currently believed. The significant correlation between the piriform and orbitofrontal cortices is an interesting finding too, although not so surprising. Finally, the significant correlation between the piriform and entorhinal cortices could be explained considering that they are both central olfactory connections.
Cortical thickness studies are not common in the literature. Feczko et al. described a novel protocol for measuring the thickness, surface area and volume of three medial temporal lobe sub-regions. Participants included younger (ages 18–30) and older (ages 66–90) normal subjects, as well as patients (ages 56–90) with mild Alzheimer’s disease (AD). Cortical surface models were reconstructed from the grey/white and grey/cerebrospinal fluid boundaries, and a hybrid visualisation approach was implemented to trace the entorhinal cortex, perirhinal cortex and parahippocampal cortex, using both orthogonal magnetic resonance imaging (MRI)—slice- and cortical surface-based visualisation of landmarks. They found that the entorhinal cortex thickness in younger normal individuals was 2.65–2.94 mm (n = 58), in older normal individuals was 2.54–2.76 mm (n = 94) and in AD patients was 2.03–2.28 mm (n = 58). The MV ± SD of our study’s E was 2.11 ± 0.87 mm (n = 37). Given the similarity of age between older normal individuals of their study and our specimens, the comparatively thinner entorhinal cortex of our specimens could be either due to the restricted point of the entorhinal cortex we chose or due to the formalin-effect to the specimen.
It is controversial whether entorhinal cortex atrophy is present in normal aging. It is possible that such effects observed in clinically normal older groups result from the presence of subclinical AD pathology. With the use of MRI data at the current resolution, it can be difficult for computational algorithms to exclude the dural layer that typically overlies the crown of the parahippocampal gyrus, potentially artificially thickening the estimate of entorhinal cortex thickness. This could be another explanation for the comparatively thinner entorhinal cortex of our specimens, and in our opinion, the most probable. We consider our anatomical method for the measurement of cortical thickness as being precise and reliable (we were able to see and touch the cortical layer).
The NA is involved in the pathophysiology of depression and is moreover a deep brain stimulation target for selected patients suffering from treatment-resistant depression. According to the findings of van Tol et al., reduced volume of the rostral-dorsal, anterior CG is a generic effect in depression and anxiety disorders, independent of illness severity, medication use and sex. This generic effect supports the notion of a shared aetiology and may reflect a common symptom dimension related to altered emotion processing. Early onset of depression is associated with a distinct neuroanatomical profile that may represent a vulnerability marker of depressive disorder.
Our study indicates that the cingulate cortex is probably the thickest cortical area connected to the NA and suggests a potentially more significant relation between the orbitofrontal cortex and the limbic lobe, than what is currently believed. Furthermore, we provided evidence that the NA size is correlated neither with the thickness of its cortical connections nor with the size of the CG. We have also provided a reliable anatomical method for cortical thickness measuring. Further research is needed to explain the correlations we found, setting them in a wider frame of the understanding of brain mechanisms. Functional brain techniques seem, in our opinion, more appropriate for this purpose.
AC, anterior commissure; AD, Alzheimer’s disease; CG, cingulate gyrus; CGH, cingulate gyrus height 12 mm rostral to the anterior commissure; MRI, magnetic resonance imaging; MV, mean value; NA, nucleus accumbens; PC, posterior commissure; SD, standard deviation; SN, substantia nigra; STN, subthalamic nucleus; VTA, ventral tegmental area.
This study was carried out in the context of the first author’s doctoral research entitled: ‘Stereotactic neurosurgical anatomy of the nucleus accumbens’ which took place at the Department of Anatomy, University of Athens School of Medicine, Athens, Greece.
All authors contributed to the conception, design, and preparation of the manuscript, as well as read and approved the final manuscript.
All authors abide by the Association for Medical Ethics (AME) ethical rules of disclosure.
Measurements of the cingulate gyrus height and thickness of the cingulate, orbitofrontal, piriform and entorhinal cortices.
|Hemisphere||CGH (mm)||C (mm)||O (mm)||P (mm)||E (mm)|
C, cingulate cortex thickness anterior to the genu of the corpus callosum; CGH, cingulate gyrus height 12 mm rostral to the anterior commissure; E, entorhinal cortex thickness at the midpoint of the hippocampal uncus; L, left; MV, mean value; n, sample number; O, orbitofrontal cortex thickness at the AC–PC plane, 10 mm laterally off the midline; P, piriform cortex thickness posterior to the olfactory tract; R, right; SD, standard deviation.