The Amygdaloid Body as the Anatomical Substrate of Emotional Memory: Implications in Health and Disease

Abstract

The Amygdaloid Body is a heterogeneous nuclear complex that establishes extensive connections with numerous structures of the limbic system, the thalamus, the brainstem, and the neocortex, and constitutes the focal center of its widespread three-dimensional white matter chassis. Since the 50s, the neurophysiological observations of Wilder Penfield et al. began to clarify the role of the AB in human memory. More recently, the introductions of a more advanced neuroimaging technology (PET, fMRI, DTI) led to a growing awareness of its crucial implications in the etiology of a variety of neuropsychiatric disorders, such as trauma spectrum and mood spectrum disorders. Additionally, the AB and its connections have been successfully used as a target for Deep Brain Stimulation (DBS) in the treatment of refractory forms of psychiatric disorders, especially trauma spectrum disorders. Therefore, gaining a deeper understanding of the morphophysiology of the AB has increasingly become utmost relevance for neuroscientists and clinicians alike. With the present chapter, we attempt to provide an exhaustive description of the functional anatomy of the AB, hopefully providing a useful tool for the approach to the anatomical substrates of the emotional components of memory and learning and to their role in the phenomenology and treatment of neuropsychiatric disorders.

1. Introduction

“Learning is defined as the change in behaviour that results from acquiring knowledge about the world, and memory as the process by which knowledge is encoded, stored, and later retrieved” [1].

Debate on the mind’s mechanism for learning and memory dates back to the beginnings of western science when the ancient Greek philosophers speculated about the causes of behavior and the relation between mind and brain. In the seventeenth century, René Descartes believed in a dualistic view between body and mind. According to his view, the brain (part of the body) is responsible for perception, motor acts, memory, appetites, and passions, while mind for the conscious experiences characteristic of the human behavior. The connection between body and mind was supposed to be in the pineal gland, located in the center of the brain. In the eighteenth century, the philosophical debate about learning and memory split along different lines: Empiricists believed that the brain was initially a blank slate (tabula-rasa) that is later filled by sensory experiences, whereas idealists, notably Immanuel Kant, believed that the perception of the world was determined by inherent features (a priori) of human mind [1].

During the twentieth century, Freud and colleagues introduced the psychoanalytical thinking, which consisted in a renewed dualistic model of memory and learning based on conscious and unconscious processes [2]. Despite the importance of this revolutionary intuition, its biological confirmation had to wait for almost one century later when the development of modern neuroscience, open to the possibility of reconsidering previous philosophical thoughts, embraced a more pragmatic model based on functional and molecular neuroanatomy [1, 3, 4, 5]. The first observation to have been confirmed was properly the intuition of Descartes that there would be a higher-order matter (soul) controlling a lower one (body): since the second half of the twentieth century, it has been assumed that phylogenetic recent areas of the brain, disposed cranially in the neuraxis, control older areas disposed more caudally. On this respect, Paul MacLean in 1960 presented the model of the “triune brain,” comprising a reptilian central primitive brain, devoted to homeostasis and survival; a paleomammalian brain, surrounding the central reptilian part, which is related to emotional and motivational aspects of behavior; and lastly, a neomammalian brain, corresponding to the neocortex, which is responsible for cognitive and executive integrations of behavior [6]. More recently, Heimer and Wilson [7] demonstrated that basal ganglia extend all the way to the ventral surface of the mammalian brain to include the olfactory tubercle and other structures of the basal proencephalon. They grouped together these portions of the basal ganglia into what they termed ventral striato-pallidal system. Such model replaced the old concept of the subcommissural substantia innominata, providing a more defined morphophysiologic establishment to these structures. The importance of this acquisition is not only due to more refined topographic depiction of the basal ganglia, but also and foremost to the understanding of the biunivocal relationship between the basal ganglia and the whole cortical mantle.

Contemporarily, the use of neural electrical stimulation in awake-neurosurgery, employed in the treatment of refractory temporal lobe epilepsy, allowed to clarify the anatomical substrate of memory and learning [1, 8, 9, 10]: it was demonstrated that mnesic functions were not grouped in a single region but widely distributed among many brain regions, whose access was independent of visual, verbal, or other sensory clues. The current classification of memory (shown in Figure 1) was elaborated by Peter Graf and Daniel Schacter, who based their model on the observation of post-operative mnesic impairments [1, 10, 11]. The authors classified mnesic functions in a short-term type and in a long-term type. Long-term memory, in turn, has been subdivided into implicit and explicit forms, differing in that the former does not require conscious awareness for recalling. Four different processes compose the physiologic build-up process of long-term memory: encoding, storage, consolidation, and retrieval [12].

Figure 1.

Classification of memory by Graf and Schacter.

Implicit memory (also referred to as non-declarative or procedural memory) is unconscious, mainly automatic and tightly connected to the original conditions under which the learning occurred. It gives rise to priming, skill learning, habit memory, and conditioning and comprises the emotional and viscerosomatic memories.

Explicit memory (or declarative memory) is defined as the deliberate or conscious retrieval of previous experiences, as well as conscious recall of factual knowledge about people, places, and things and it can be further subdivided into an episodic or autobiographical memory (memory of personal experiences) and a semantic memory (memory of meanings, words, and concepts).

But why do humans store some memories and discard others? This question has persisted since the beginning of the philosophical debate on memory and learning. Nowadays, the intuitive assumption that memory consolidation and storage depend, to a substantial extent, on the affective charge of experiences is well established. Such function is largely under the control of the Amygdaloid Body (AB) [13]. Indeed, the AB signals to the hippocampus the experiences that are more emotionally salient and should have a sort of priority for storage, relative to other poorly salient experiences that are less significant for the individual. As a consequence, the AB is thought to be involved in the genesis of the affective symptomatology of many neuropsychiatric disorders, such as trauma spectrum and mood spectrum disorders, in which the emotional features of mnesic functions are often altered [14, 15, 16, 17]. Accordingly, the AB and the connections of its functional network have been successfully used as a target for Deep Brain Stimulation (DBS) in refractory cases of some of these conditions [18, 19, 20, 21, 22, 23, 24, 25, 26].

In this scenario, gaining a deeper understanding of the morphophysiology of the AB has become increasingly relevant not only for neuroscience researchers but also for the clinicians who, in their daily practice, deal with the aberrations of such fascinating aspects of neurobiology.

2. Neuroanatomy

The AB is located deep within the anterior segment of the uncus and it appears as an ovoid-shaped gray matter nuclear complex with an anteroposterior and transverse diameter of 12 mm and of 16 mm, respectively. The medial surface of the AB is in relation to the semilunar and ambiens gyri. Laterally, the posterior aspect of the AB faces the intralimbic gyrus and piriform cortex, (Figure 2a) and medially constitutes the anterolateral wall of the temporal horn of the lateral ventricle. This surface is concave due to the presence of the head of the hippocampus, from which it is separated by the uncal recess (Figure 2b). In the temporal horn, the superior portion of the AB is within the anterior roof of the ventricle, and anteriorly within the white matter of the temporal pole (Figure 2b). Superiorly, the AB is continuous with the globus pallidus and with the ventral claustrum (Figure 3) [27].

Figure 2.

Morphology and location of the amygdaloid body (*). In “a”, exposure of the medial temporal region is shown. The amygdaloid body is located deep within the anterior segment of the uncus. The relation between the amygdaloid body and the periamygdaloid cortex is highlighted in green. The medial surface of the amygdaloid body is adjacent to the semilunar gyrus, superiorly, and the ambiens gyrus, inferiorly. Postero-medially and postero-inferiorly, the amygdaloid body faces the intralimbic gyrus and the piriform cortex, respectively. In “b”, the dissection of the basal aspect of the left temporal lobe is shown. The relation between the amygdaloid body and the surrounding structures is highlighted. The uncus consists of an anterior portion, containing the amygdaloid body and covered by the peri-amygdaloid cortex, and a posterior portion, enclosing the head of the hippocampus and covered by the cortex of the parahippocampal gyrus. The amygdaloid body is separated from the head of the hippocampus by the uncal recess. Anteromedially, the amygdaloid body is separated from the anterior perforated substance by the rhinal sulcus.

Figure 3.

Coronal cut of the medial temporal lobe at the level of the anterior commissure showing the nuclear subdivision of the amygdaloid body, in macroscopic direct-view. In “a”, the amygdaloid nuclei are highlighted. The basolateral amygdaloid body (yellow dotted line) is composed of the lateral (blue), basolateral (red), and basomedial (yellow) nuclei, disposed in a latero-medial direction. The lateral nucleus exhibits a striate appearance, due to the passage of fibers of the ventromedial portion of the uncinate fascicle. The centromedial amygdaloid body (blue dotted line) is composed of the central nucleus (purple), located infero-laterally and anteriorly, and the medial nucleus (pink), located superomedially and posteriorly. The cortical amygdaloid body (green) comprises multiple tiny nuclei located superomedially, in the context of the nuclear complex. In “b”, the relation between the amygdaloid body and the surrounding structure is highlighted. The ventral claustrum is scattered throughout the uncinate fascicle, without a clear demarcation from the basolateral amygdaloid body. The amygdalo-piriform area is composed of the gray matter located between the amygdaloid body and the piriform cortex. Superiorly, the amygdaloid body adjoins the white matter of the basal forebrain and it is continuous cranially with the globus pallidus.

2.1 Microneuroanatomy

Burdach was the first, in the early nineteenth century, to provide a comprehensive description of this structure, for which he proposed the name “amygdala” in view of its resemblance to an almond seed [28, 29, 30, 31]. During the last century, several authors have improved the anatomical knowledge of the region, identifying a large number of heterogeneous nuclei surrounding the one originally described by Burdach. Accordingly, the name of this area of gray matter was changed to the more appropriate term AB [31].

This nuclear complex is currently subdivided into four groups of nuclei: a basolateral complex, which is the one originally described by Burdach, a centromedial complex, a cortical complex, and an additional fourth group [27, 31], as shown in Figure 3.

The basolateral AB has a telencephalic origin and is composed by the lateral, basolateral, and basomedial nuclei. The lateral nucleus exhibits a striate appearance, due to the presence of fibers of the ventromedial uncinate fascicle.

The centromedial AB has a diencephalic origin and is composed of the central nucleus, located inferolaterally and anteriorly, and the medial nucleus, located superomedially and posteriorly, adjacent to the stria terminalis. Both nuclei are pale, reminding the color of the pallidum with which they share the ontogenetic origin.

The cortical AB has a telencephalic origin and is composed by multiple tiny nuclei located superomedially, in the context of the nuclear complex. These nuclei are continuous with the piriform cortex of the uncus, and include the primary and secondary nuclei of the lateral olfactory stria and the anterior cortical amygdaloid nucleus intercalated in the olfactory system.

The additional group comprises all gray matter areas surrounding the proper nuclear complex, and seamlessly transitions into the neighboring structures. It includes two essentially independent structures: the telencephalic amygdalo-piriform area, located between the AB and the piriform cortex; and the diencephalic bed nucleus of the stria terminalis, which identifies with the gray matter following the course of the stria terminalis and constitutes the continuity between the AB and the hypothalamus.

2.2 Connectomic

The AB is characterized by numerous connections with the limbic system, the thalamus, the brainstem, and the neocortex, organized in a three-dimensional chassis around the nuclear complex [12, 30].

2.2.1 The amygdalofugal pathways

Humans and other mammals are endowed with two divergent amygdaloid projection systems named dorsal and ventral amygdalofugal pathways [30, 31, 32, 33]. These two bundles arise from the centromedial AB and from the basolateral AB, respectively, and converge at the level of the septum and of the hypothalamus [30].

The dorsal amygdalofugal pathway, shown in Figure 4, forms the stria terminalis. This projection system has been comprehensively described in humans [27, 30, 31] as a C-shaped bundle arising from the centromedial AB and following the lateral ventricle and the surface of the thalamus toward the ipsilateral foramen of Monro. Hence, it splits into three components: a pre-commissural component, reaching the septum and the nucleus accumbens septi; a commissural component, running toward the contralateral hemisphere within the anterior commissure (AC) and taking part in the inter-amygdaloid pathway; and a post-commissural component running behind the AC. The latter, splits into a portion following the post-commissural fornix to reach the mammillary body, and another portion following the stria medullaris thalami to reach the habenula, the thalamus, and the hypothalamus. All along its course, the stria terminalis is in continuity with its bed nucleus. The bed nucleus of the stria terminalis is subdivided into a rostral, intermediate, and caudal portions and it represents the anatomical gray matter continuity between the AB and the hypothalamus.

Figure 4.

Dissection of the medial aspect of the right hemisphere. The complete course of the stria terminalis is exposed after the complete removal of the fornix and the hypothalamus (which position is still marked in green). The stria terminalis is a C-shaped bundle arising from the centromedial amygdaloid body (*) and following the lateral ventricle and the surface of the thalamus toward the ipsilateral foramen of Monro, where it splits into three components: A pre-commissural component, reaching the septum and the nucleus accumbens septi; a commissural component, running toward the contralateral hemisphere within the anterior commissure and taking part of the inter-amygdaloid pathway; and a post-commissural component running behind the anterior commissure. The latter splits in a portion following the post-commissural fornix and another portion following the stria medullaris thalami to reach the mammillary body, the habenula, the thalamus, and the hypothalamus.

The ventral amygdalofugal pathway arises from basolateral AB and runs ventrally to the ventral striato-pallidal structures, the AC, and the internal capsule, to reach the septal region. Due to the spread diffusion of its fibers into several structures and due the limited size of the region, a comprehensive description of its anatomy has always been difficult to achieve [27, 30, 31]. Using Klinger’s fiber dissection technique on ex-vivo brain specimens, we have recently provided a new account on the subject [27]. According to our findings, the ventral amygdalofugal pathway appears as a widespread bundle connecting the basolateral AB with the septum, the thalamus and the hypothalamus, characterized by two arms diverging below the AC, as shown in Figure 5.

Figure 5.

In “a” and in “b”, the dissection of the medial and inferior aspects of the same right hemisphere are highlighted, respectively. Both pictures show the complete course of the ventral amygdalofugal pathway, which arises from the basolateral amygdaloid body (*) and immediately spreads into two arms. The anterior arm of the ventral amygdalofugal pathway, named also diagonal band of Broca, runs underneath the anterior commissure in a latero-medial direction toward the basal nucleus of Meynert (scattered along its course) and bends anteriorly and medially toward the nucleus accumbens septi, directed to the septum. The posterior arm of the ventral amygdalofugal pathway runs laterally above the anterior perforated substance and toward the thalamus and hypothalamus, where it intermingles with fibers of the medial forebrain bundle and of the stria medullaris thalami.

The anterior arm of the ventral amygdalofugal pathway, named also diagonal band of Broca, runs underneath the AC in a latero-medial direction toward the basal nucleus of Meynert (scattered along its course) and bends anteriorly and medially toward the nucleus accumbens septi, directed to the septum. The posterior arm of the ventral amygdalofugal pathway runs laterally above the anterior perforated substance and toward the thalamus and hypothalamus, where it intermingles with fibers of the medial forebrain bundle (MFB) and of the stria medullaris thalami.

2.2.2 Connections with the brainstem

The MFB connects the AB with the brainstem [4, 34], as shown in Figure 5, and is supposed to play a role in the neurovegetative correlates of implicit memories. Since the rostral portion of the MFB splits into several different branches, extending in various directions and depths, its dissection in humans has always been partial [30, 35, 36, 37, 38]. In the diffusion MRI tractographic study performed by Coenen et al. [39] the MFB appeared composed by a main trunk arising in the dentate nucleus of the cerebellum, and running through the retrobulbar area and the periaqueductal gray, to reach the ventral-tegmental area of the midbrain. Here, it splits into an inferomedial portion, running along the lateral wall of the third ventricle and reaching the hypothalamus; and in a superolateral portion, running inferolaterally to the thalamus. In particular, the superolateral MFB has been dissected in ex-vivo specimens, from the ventral-tegmental area, caudally, to the medial ceiling of the anterior perforated substance, rostrally, where it intermingles with the posterior portion of the ventral amygdalofugal pathway [27]. Its neuromodulation in humans seems to show mood-regulating effects in refractory major depressive disorder (MDD) [25].

2.2.3 Connections with the hippocampus

The intra-ventricular amygdalo-hippocampal bundle, shown in Figure 6, connects the basolateral AB with the head of the hippocampus. This connection was firstly described in rhesus monkeys by Rosene and Van Hoesen in 1977 [40], and in rats by Kemppanien et al. in 2002 [41]. In humans, it was identified by Di Marino et al. in 2016 [31], and by our group in 2021 [27]. Kemppainen and Pitkanen [42] demonstrated that the connections between the basolateral AB and the CA1/subiculum of the hippocampus are resistant to neuronal damage induced by the status epilepticus in rats, suggesting their role as a pathway for seizure propagation between the two structures. Di Marino et al. [31] suggested the existence of an additional extra-ventricular amygdalo-hippocampal connection, whose existence has still to be confirmed [27, 31].

Figure 6.

In “a”, the lateral dissection of the right hemisphere is shown, while in “b” and in “c”, its stepwise basal dissection is shown. In “a”, the removal of the inferior longitudinal fascicle allows the exposure of the amygdaloid body (*), anteriorly and of the hippocampus posteriorly. Retracting the hippocampus postero-inferiorly with a spatula was possible to expose the tiny bundle that connects the aforementioned structures. Such connection is exposed also in “b”, where the cingulum has been partially sectioned and the hippocampus retracted medially (exposing the roof of the temporal horn of the lateral ventricle). “c” Highlights the amygdalo-temporal fascicle.

2.2.4 Cortical connections

The amygdalo-temporal fascicle, whose function is still unknown, arises from the basolateral AB and runs straight to the cortex of the temporal pole, as shown in Figure 6. Its existence in humans was first mentioned by Curran in 1909 [43], but it has been described in detail only by Klingler and Gloor in 1960 [30] and by our group in 2021 [27].

The temporo-pulvinar bundle of Arnold arises from the anterior temporal cortex with a postero-medial course and is arranged in a thin sheet of fibers running above the roof of the temporal horn underneath the tail of the caudate nucleus, establishing a connection with the lateral AB. Afterwards, it continues posteriorly within the sublenticular segment of the internal capsule, eventually ending into the pulvinar. Its dissection is shown in Figure 7. Such bundle was first detected by von Monakow in 1895 [44] and extensively described by Klingler and Gloor [30], who termed it “inferior thalamic peduncle,” and by our group [27]. Its function is still unknown, but clinical evidences of neuromodulation suggest that it could play a role in mood control [45].

Figure 7.

Dissection of the inferior aspect of the right hemisphere. “a” Highlights the relation between the amygdaloid body (*) and the uncinate fascicle, located anterolaterally, the lateral olfactory stria, located anteromedially, and the temporo-pulvinar bundle of Arnold located postero-laterally. The lateral olfactory stria arises from the olfactory tract, in front of the anterior perforated substance and runs laterally, just anteriorly to the diagonal band, to intermingle with fibers of the external capsule at the level of the limen insulae and insular gyri. Laterally, the lateral olfactory stria provides fibers directed toward the cortical amygdaloid body. The temporo-pulvinar bundle of Arnold arises from the anterior temporal cortex with a postero-medial course and is arranged in a thin sheet of fibers running above the roof of the temporal horn underneath the tail of the caudate nucleus, establishing a connection with the lateral amygdaloid body. Afterwards, it continues posteriorly within the sublenticular segment of the internal capsule, eventually ending into the pulvinar. In “b”, the lateral olfactory stria has been sectioned exposing the anterior commissure.

The lateral olfactory stria, whose dissection is shown in Figure 7a, arises from the olfactory tract, in front of the anterior perforated substance, and represents the lateral part of the olfactory trigon. It runs laterally, just anteriorly to the diagonal band, to intermingle with fibers of the external capsule at the level of the limen insulae and insular gyri. Laterally, the lateral olfactory stria provides fibers directed toward the cortical AB, in particular, it reaches the primary and secondary olfactory nuclei and the anterior cortical amygdaloid nucleus. Along its course, the lateral olfactory stria interconnects the olfactory bulbs, the gray matter of the anterior perforated substance (considered the involution of the laminated-olfactory tubercle of macrosmatic mammals), the piriform cortex, the AB, and the insula; for this reason, Nieuwenhuys et al. [4] have considered such pathway as the anatomical substrate of the link between taste and smell, as well as between smell, emotions, and memories. It is widely ascertained that olfactory clues are the strongest inputs for perceptual recalling of emotional components of memories (commonly referred to by the French as Proust phenomenon) [46]. In fact, the lateral olfactory stria connects the olfactory bulb directly with the AB and the piriform cortex, making olfaction the sole sensory modality to reach the cortex without thalamic retransmission.

Herrick in 1956 wrote: “This olfactory field at the anterior end of the brain is the dominant centre of control of all behaviours of these primitive vertebrates, and for this reason it was the seedbed for further structural differentiation as the patterns of behaviour were stepped up from one integrative level to another. Here, the rudimentary cortex had its beginnings” [5].

The uncinate fascicle (UF), whose dissection is shown in Figure 7 and in Figure 8, connects the frontal lobe with the anterior portion of the temporal lobe. It is composed by a ventromedial part, arising from the medial orbitofrontal cortex (mOFC), and an anterolateral part, arising from the medial prefrontal cortex (mPFC). In its course to the temporal pole, the UF bends below the nucleus accumbens septi forming the supero-anterior wall of the Gratiolet’s canal, where some fibers of its ventromedial part provide connections with the basolateral AB, as shown in Figure 8c. The ventral claustrum is scattered throughout the ventromedial UF, without a clear demarcation from the AB. The fibers of the anterolateral UF follow their course straight to the cortex of the temporal pole [27, 30].

Figure 8.

The dissection of the lateral aspect of the left hemisphere. In “a”, the whole course of the uncinate fascicle that connects the frontal and the temporal lobes is shown. Its relation with the ventral claustrum, scattered throughout its course without a clear demarcation from the amygdaloid body (*), is highlighted in “b”. The uncinate fascicle is composed by a ventromedial part, arising from the medial orbitofrontal cortex, and an anterolateral part, arising from the medial prefrontal cortex. In its course to the temporal pole, it bends below the nucleus accumbens septi forming the supero-anterior wall of the Gratiolet’s canal, and, at this point, some fibers of its ventromedial part provide connections with the basolateral amygdaloid body as shown in “c”.

The most conspicuous AB’s cortical connection is constituted by the cingulum, whose dissection is shown in Figure 9. The cingulum, in humans, arises from the basolateral AB and runs over the medial surface of the hippocampus in a postero-medial direction (as shown in Figure 9b) interconnecting the gyrus ambiens and the parahippocampal gyrus. Afterwards, it bends anteriorly, above the corpus callosum, reaching first the parietal cortex, then the mPFC, eventually ending at the level of the basal forebrain [27]. Since the observation of Dejerine in 1895, there has been the idea that the cingulum would be composed by the combination of short fibers, arising from neighboring gyri [32, 47]. This observation was confirmed in 2014 by the tract-tracing study performed by Heilbronner and Haber [48], and in 2021 by our dissection work [27], which illustrated fibers arising from the basolateral AB getting directly into the temporal cingulum, and indirectly, through the ventral amygdalofugal pathway, into the rostral-subgenual cingulum on the medial surface of the nucleus accumbens septi.

Figure 9.

The dissection of the medial aspect of the right hemisphere. In “a”, the whole course of the cingulum is shown, which arises from the basolateral amygdaloid body (*), and runs over the medial surface of the hippocampus in a postero-medial direction interconnecting the gyrus ambiens and the parahippocampal gyrus. Afterwards, it bends anteriorly, above the corpus callosum, reaching first the parietal cortex, then the medial prefrontal cortex, eventually ending at the level of the basal forebrain. The arising point of the cingulum from the basolateral surface of the amygdaloid body is shown in “b”.

The ABs are reciprocally connected by the AC, the major inter-hemispheric connection of the limbic system. The AC, which course is shown in the dissection in Figure 10, is a transverse, handlebar-shaped bundle composed by a compacted medial portion that spreads bilaterally into the temporal lobes. The medial AC is located in front of the columns of the fornix, at the rostral edge of the third ventricle, and is housed into a canal formed entirely by white matter, named canal of Gratiolet. Such canal is composed by the UF, anteriorly and superiorly, from the diagonal band of Broca anteriorly and inferiorly, and from the columns of the fornix, posteriorly. From its medial portion, the AC curves bilaterally and posteriorly passing under the lenticular nucleus before crossing the temporal isthmus. At this point, it splits into an anterior crus extending forword to the mPFC, and in a fan-shaped posterior crus giving branches getting into the AB’s dorsal surface [27, 30, 49]. Eventually, the posterior crus divides into an occipital extension, joining the sagittal stratum, and a temporal extension, reaching the anterolateral-temporal cortex.

Figure 10.

The dissection of the lateral aspect of the left hemisphere in “a” and of the inferior aspect of the brain in “b” and “c” show the complete course of the anterior commissure starting from its compacted medial part housed into the Gratiolet’s canal and spreading bilaterally into and anterior and a posterior course reaching the frontal and temporal lobes and interconnecting the two amygdaloid body.

2.3 Functional anatomy of the AB

The heterogeneous pattern of projections of the AB has been recently organized in two different functional systems by de Olmos and Heimer [50]: the extended amygdala and the great limbic lobe.

The extended amygdala belongs to the ventral pallidum and is constituted by a continuum of gray matter of diencephalic origin organized as a ring encircling the internal capsule. It is composed by the centromedial AB, the bed nucleus of the stria terminalis, and the lateral and medial hypothalamus. All structures belonging to the extended amygdala are shown in the dissection in Figure 11. These structures are reciprocally connected through the stria terminalis, which is supposed to link food-related, olfactory, sexual, and endocrine cues [12, 50]. In particular, the central AB seems to be directly connected to the autonomic and somatomotor centers in the lateral hypothalamus, while the medial AB seems to be directly connected to the endocrine-related medial hypothalamus.

Figure 11.

Shows the structures composing the extended amygdala that belongs to the ventral pallidum and is constituted by a continuum of gray matter of diencephalic origin organized as a ring encircling the internal capsule. It is composed by the centromedial amygdaloid body (*), the bed nucleus of the stria terminalis (light blue) and the lateral and medial hypothalamus (green). These structures are reciprocally connected through the stria terminalis (red dashed line).

The great limbic lobe, shown in Figure 12, includes the basolateral AB, the cortical AB, the hippocampus, the basal forebrain, the insula, the cingulate gyrus, the ventral prefrontal cortices, the subcallosal, and the entorhinal cortices. All these structures have a telencephalic origin, and its subcortical nuclei belong to the ventral striatum. The white matter bundles connecting the structures of the great limbic lobe are the ventral amygdalofugal pathway, the cingulum, the UF, the MFB, and the lateral olfactory stria [12, 50].

Figure 12.

Shows the brain cortical regions composing the great limbic lobe that includes the basolateral and cortical amygdaloid body (*), the hippocampus, the basal forebrain, the insula, the cingulate gyrus, the prefrontal cortices, and the subcallosal, entorhinal, and ventral-tegmental areas.

The converging point between the extended amygdala and the great limbic lobe is the basolateral AB, that receives afferent fibers from the cortex and from the ventral striatum, and projects efferent fibers to the centromedial AB.

In view of the recent advances in functional neurosurgery, such classification acquires a clinical relevance. Indeed, the basolateral AB has been used as DBS-target for the treatment of refractory Post-Traumatic Stress Disorder (PTSD), [22, 23, 26] autism, [19, 20, 21] and, together with the cingulum and the superolateral portion of the MFB, refractory MDD [18, 19, 20, 51]. Furthermore, DBS targeting the bed nucleus of the stria terminalis has been shown effective in decreasing anxiety in patients with obsessive-compulsive disorder (DOC) [24] and relapsing anorexia [25]. These observations are in line with the results of analogous studies in animals [52]. On the other hand, Piacentiani et al. [53] reported that the unintentional stimulation of the stria terminalis, caused by the displacement of a DBS-electrode previously positioned in a patient affected by dystonia, determined the development of MDD with mood-incongruent psychotic symptoms. These pieces of evidence support the idea that the stimulation of the AB’s connections, comprised in the great limbic lobe, might increase mood control and decrease the severity of aversive emotional responses to traumatic memories [20, 21, 22, 23, 25, 26, 51, 54, 55, 56, 57]. On the contrary, the effects of the stimulation of the extended amygdala in humans are still controversial and the few case reports in the literature are not enough to predict its possible outcomes [24, 25, 53]. All results are resumed in Table 1.

Paper	Year	Target	Disorder	Effect
Kennedy et al. [18]	2011	Basolateral amygdaloid body	Autism	Improve
Langevin et al. [19]	2012	Basolateral amygdaloid body	Autism	Improve
Langevin et al. [19]	2012	Cingulum	MDD	Improve
Schlaepfer et al. [20]	2013	Basolateral amygdaloid body	Autism	Improve
Schlaepfer et al. [20]	2013	Basolateral amygdaloid body	MDD	Improve
Sturm et al. [21]	2013	Superolat. Portion of the MFB	MDD	Improve
Koek et al. [22]	2014	Basolateral amygdaloid body	PTSD	Improve
Langevin et al. [23]	2015	Basolateral amygdaloid body	PTSD	Improve
Blomstedt et al. [25]	2017	Medial forebrain bundle	MDD	Improve
Lavano et al. [26]	2018	Basolateral amygdaloid body	PTSD	Improve
Piacentiani et al. [53]	2008	Stria terminalis (unintentional)	Dystonia	Developing of MDD with psychosis
Luyten et al. [24]	2016	Bed nucleus of the stria terminalis	DOC	Improve
Blomstedt et al. [25]	2017	Bed nucleus of the stria terminalis	Relapsing anorexia	Improve

Table 1.

Summary of the clinical outcomes in patients treated with the DBS of the amygdaloid body and its connections.

3. Role of the AB in trauma spectrum disorders

Trauma spectrum disorders are a group of psychological alterations consequent to the exposure to a trauma, defined as an event that is perceived to be life-threatening or to pose the potential of serious bodily injury to self or others [58]. This condition is considered as a chronic dysregulation of the physiologic nervous response that normally follows the exposure to a threat.

Trauma spectrum symptomatology is characterized by phenomena that can be grouped into three primary domains: reminders of the exposure such as flashbacks, intrusive thoughts, and nightmares; activation patterns such as hyperarousal, insomnia, agitation, irritability, impulsivity, and anger; and deactivation patterns such as numbing, avoidance, withdrawal, confusion, derealization, depression, and dissociation. In particular, dissociation is defined as a failure, disruption, interruption, and/or discontinuity of the normal, subjective integration of behavior, memory, identity, consciousness, emotion, perception, body representation, and motor control, and it represents the link between trauma spectrum disorders [58] and a large variety of psychiatric disorders like drug abuse, self-injurious behaviors, suicidality, somatization, and borderline personality disorder [59, 60].

Dissociative disorders can be classified into two subtypes: structural dissociation, involving the autonomic responses following the exposure of traumatic cues, leading to possible repercussion on personality and general functioning; and somatoform dissociation, which is characterized by the expression of implicit viscerosomatic memories, leading to body representation disorders and to a large variety of dermatologic and immune disorders [59, 60].

Clinical manifestation of the psychological trauma has been known since the beginning of the psychoanalysis [2], although the deep insight of its consequences on the society was reached only in the second part of the nineteenth century when it has been given a bigger attention to war veterans difficulties at the moment of their return to daily living. The huge increase of the post-traumatic diagnoses in the last decades, whose recognition has moved far beyond the border of war situations, along with their heavy potential for social impairment, highlighted the need for a more accurate nosology and a deeper insight into the pathophysiology of these disorders.

It is currently thought that there is a substantial continuity in the expression of trauma spectrum disorders and, accordingly, trauma manifestations have been organized in two groups of clinical severity: the simple and the complex PTSDs [58, 59, 60].

Simple PTSDs are characterized by the presence of a single or multiple traumatic event/s that occurred after the psychological maturity of the subject. These individuals typically retain both explicit and implicit memories of the trauma. Simple PTSDs comprise the acute post-traumatic disorder and the proper PTSD.

Complex PTSDs are characterized by the presence of a single or more often multiple relational traumatic event/s that occurred before the psychological maturity of the subject. Individuals with Complex PTSD typically retain implicit memories of the traumatic event(s), but not necessarily explicit memories. Complex PTDSs comprise the proper complex PTSD, the Dissociative Amnesias, the depersonalization/derealization disorders, and the dissociative identity disorder.

This classification has been supported by functional magnetic resonance studies, which have observed differences in the cerebral activation pattern between the simple and complex PTSDs [59, 60, 61]. According to these studies, when exposed to traumatic cues, a subject affected by a simple PTSD presents a hyperactivation of the AB, a relative hyperactivation of the autonomic nervous system (ANS), and a deactivation of the mPFC; differently, in a subject affected by a complex PTSD, the activation pattern might not be so well defined and a hyperactivation of the AB/ANS might not be so evident, the thalamus might be deactivated, and some parts of the cortex might present a different degree of activation/deactivation.

These neurobiological evidences found clinical confirmation in the symptomatic expression of the two groups of post-traumatic disorders: in simple PTSDs, the symptoms consequent to the reminder of the exposure are prevalent, while in complex PTSDs the symptomatic pattern is more characterized by hyperactivation and deactivation [58, 62].

In Table 2 the biological evidences reported in the PTDS are resumed. The endocrine system presents activation of the hypothalamic-pituitary-adrenal axis and of the hypothalamic-pituitary-thyroid axis. The cerebral neurochemical pattern is characterized by increased levels of noradrenalin and by decreased levels of serotonin. Eventually, some parts of the brain like the hippocampus, the mPFC, and the anterior cingulate gyrus present a volume reduction [62].

Feature	Change	Effect
1. Neuroendocrine
Hypothalamic-pituitary-adrenal axis	Hypocortisolism	Disinhibits CRH/NE and upregulates response to stress; Drives abnormal stress encoding and fear processing
	Sustained, increased level of CRH	Blunts ACTH response to CRH stimulation; Promotes hippocampal atrophy
Hypothalamic-pituitary-thyroid axis	Abnormal T3:T4 ratio	Increases subjective anxiety
2. Neurochemical
Dopamine	Increased levels	Interferes with fear conditioning by mesolimbic system
Norepinephrine	Increased levels	Increases arousal, startle response, encoding of fear memories
Serotonin	Decreased levels in dorsal and median raphe	Disturbs dynamic between amygdala and hippocampus; Compromises anxiolytic effects; Increases vigilance, startle, impulsivity, and memory intrusions
GABA	Decreased levels	Compromises anxiolytic effects
Glutammate	Increased levels	Derealization and dissociation
NPY	Decreased levels	Leaves CRH/NE unopposed and upregulates response to stress
β-endorphin	Increased levels	Numbing, stress-induced analgesia, and dissociation
3. Neuroanatomical
Hippocampus	Reduced volume and activity	Alters stress responses and extinction
Amygdaloid body	Increased activity	Promotes hypervigilance and impairs discrimination of threat
Prefrontal cortex	Reduced volume and activation	Dysregulates executive functions
Anterior cingulate cortex	Reduced volume	Impairs the extinction of fear responses

Table 2.

Biological evidences reported in patients affected by PTDS.

This piece of knowledge has paved the way for the development of numerous theories regarding the pathophysiology of each AB’s connection, opening up opportunities for their clinical experimental use as DBS-targets in neuromodulation [23, 27, 62, 63].

In particular, the UF and the cingulum seem involved in the process that links implicit and explicit memories. A conduction disorder of these bundles seems to be involved in the genesis of the symptoms consequent to the reminders of the exposure and their neuromodulation showed promising results in the treatment of refractory-PTSD and MDD [19, 59].

The connections between the basolateral AB and the thalamus, such as the temporo-pulvinar bundle of Arnold and, especially, the ventral amygdalofugal pathway, seem to play a role in thalamic deactivation, and, consequently, in the development of the symptomatic deactivation pattern of PTSD [59, 61, 64].

The MFB, connecting the basolateral AB directly with the brainstem, and the stria terminalis, connecting the centromedial AB with the lateral hypothalamus could constitute a direct and an indirect links between the exposure of traumatic cues and the activation of the ANS, whose nuclei are located in the brainstem. Some authors [6, 60, 64] supported the idea that these pathways could be involved in the genesis of structural dissociation.

Additionally, the portion of the stria terminalis, connecting the centromedial AB with the medial hypothalamus, seems to be involved also in the post-traumatic endocrine dysregulation and, together with the temporo-insular connections (involved in visceral introjection belonging to the vagus nerve) could play a role in the development of the somatoform dissociation [65].

4. Role of the AB in mood disorders

The neuroanatomical and neurofunctional studies related to mood disorders are much compounded by the Leonhardian dichotomy, adopted by the DSM tradition, which categorically distinguishes between bipolar and unipolar mood disorders. As a matter of fact, the literature about the structural and functional brain correlates of both nosological groups does not yield significantly different results [66, 67] although recent imaging studies claim to have found subtle pattern differentiations [68, 69, 70]. One of the most consistent findings, when we examine the role of the amygdaloid body in mood disorders, is the tendency toward an overactivity and progressive atrophy of this brain region, particularly on the left side [66, 67, 71], but these alterations are to be mindfully considered in light of the astounding heterogeneity that characterizes mood disorders. It is possible that these amygdalar dysfunctions are not related to a specific disease, being in fact transcategorial biomarkers of dysphoric experiences, such as hyperarousal, anxiety, and irritability [71]. It is also presumable that the amygdaloid body could be more implicated with some subsets/endophenotypes of mood disorders, while being utterly marginal in other pathotypes [71]. Finally, it is likely that some changes might be state-specific (mania, hypomania, mixed states, depression), whereas other changes may be trait-related and, therefore, underlying all mood states [72]. Subsequently, while waiting for more neurobiologically-based psychopathological dimensions, addressing amygdala’s implications in affective disorders as a whole heterogeneous group seems to be the most convenient approach.

The first sound experimental contributions concerning the role of the amygdaloid body in manic-depressive illness are to be ascribed to the research of Post and coworkers [73, 74]. Largely drawing from observations of progressive epileptic threshold lowering in murine models of temporal lobe epilepsy, these authors prompted that overexcitability in the amygdaloid region could be engendered by long-term repetition of subthreshold stimuli (e.g. minor life events or subclinical mood states), at least in subjects with a substantial degree of constitutional (familial) predisposition. According to this perspective, amygdala-kindled seizures would represent a non-homologous model for mood disorders recurrence and mood episodes could be heuristically interpreted as “prolonged non-convulsive limbic pseudo-epilepsies.” Post’s theoretical model is partly prescient of later acquisitions and was historically crucial for the consolidation of the concept of mood-stabilization with antiepileptic drugs [75, 76]. A more thorough neuroanatomical model of BD was proposed by Stephen Strakowski et al. in 2012 [77]. According to their synthesis, mostly based on brain imaging studies, amygdala’s dysregulation could be placed at the center of two major emotion-regulatory cortico-striato-pallido-thalamo-cortical networks: the ventrolateral prefrontal pathway, linked to explicit and conscious emotional regulation, and the orbitofrontal pathway, associated with implicit and non-conscious affective processing [77]. This model describes essentially a top-down/bottom-up imbalance hinged upon fronto-limbic structures: a neurodevelopmental fragility in top-down regulation of the amygdaloid body and related structures, caused by an ineffective maturation of ventral prefrontal cortices or related white matter systems, would predispose subjects to extreme affective states, wherein subcortical limbic centers are insufficiently modulated and are thus free to provoke a sort of bottom-up, emotion-laden neural flooding.

Morphometric studies report contradictory findings at a first glance, showing either reductions or increases in amygdalar volumes among subjects with BD [67]. These results are to be understood with some cautiousness, since the adults with BD recruited in these studies might conceal a host of confounding factors, including differences in disease cyclicity and polarity or the differential effects of different treatment options (such as lithium salts and valproic acid). Indeed, studies investigating the issue in pediatric populations consistently show volumetric amygdalar decreases and so do research papers on first episode BD [67]. As we have proposed elsewhere, [78] molecular and cellular desensitization to glutamate as a tentative homeostatic adaptation to increased glutamatergic transmission could explain most of the findings: a chronic hyperactivity of amygdaloid structures, be it related to childhood stress or the allostatic load of mood episodes, could lead to neurotoxicity and glial cell loss. Conversely, mood stabilizers could exert neuroprotective effects on limbic structures, reducing overexcitation and normalizing receptor sensitivity. This interpretative strategy could partly resolve the outwardly contradictions found in the available scientific literature on this topic.

5. Conclusions

The relationship between the clinical expression of a disorder and its anatomofunctional correlates remains pivotal for a real understanding of the complex relations between brain and behavior [79, 80]. The disconnection paradigm, as envisaged by Geschwind in his landmark paper and revitalized today by the availability of methods for mapping connections and cortical activities in-vivo, seems to be of key importance to elaborate a comprehensive approach to the study of the brain-behavior relation. All the evidences reported in this chapter suggest that social behavior, learning, and memory depend on the activity of different brain cortical regions coordinated by subcortical structures, and that “emotions,” whose effector nuclear complex is the AB, are the key starting points for all of these physiological processes. Concurrently, there have been numerous pieces of evidence that neuropsychiatric disorders related with mnesic functions, such as trauma spectrum and mood spectrum disorders, subtend an altered connectivity linked to an abnormal activity of the AB, [4, 12, 13, 14, 15, 16, 17, 50, 51, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 77] and that its normalization by either pharmacological or non-pharmacological techniques could feasibly lead to a better clinical outcome [18, 19, 20, 21, 22, 23, 24, 25, 26, 45, 52, 54, 55, 56, 57, 68, 74, 75, 76, 78]. In this scenario, the disconnection paradigm provides a renewed mindset to rethink neuropsychiatry and psychotherapy.