This chapter describes the structure and connections of the amygdala, emphasizing three of its main components: the basolateral complex, the central nucleus and the intercalated cell masses. As we shall see in detail below, the basolateral complex is a cortex-like structure that receives most sensory inputs to the amygdala and projects to the central nucleus. In turn, the central nucleus contributes most amygdala projections to brainstem nuclei generating components of fear responses. However, because it would be would be maladaptive if all sensory inputs triggered fear, impulse traffic between the basolateral complex and central nucleus must be flexibly gated as a function of the particular combination of sensory inputs and environmental signals confronting the organism. This is the function of the intercalated cell masses, clusters of GABAergic neurons that receive glutamatergic inputs from the basolateral complex and project to central nucleus. This chapter will describe the connectivity and physiological properties underlying this process. Their system-level consequences will be considered later in this book (Chapter 9).

Is the amygdala a valid anatomical concept?

The amygdala is a nucleated structure located in the depth of the temporal lobe (Figures 3.1 and 3.2). Early anatomists divided the amygdala into three groups of nuclei (Johnston, 1923; Humphrey, 1936; Fox, 1940; Crosby & Humphrey, 1941): (1) the basolateral complex, comprised of the lateral, basolateral and basomedial nuclei; (2) the corticomedial group including the central, medial and cortical nuclei; and (3) an ill-defined anterior group including the anterior amygdaloid area, the nucleus of the lateral olfactory tract and the intercalated cell masses.

We now discuss the neuronal processes that account for the increased responsiveness of brainstem, thalamic and cortical neurons during wakefulness and REM sleep. The closing of thalamic gates during slow-wave sleep (see Chapter 6) is modified upon transition to both brain-active states under the influence of increased neuronal activities in ascending activating modulatory (mainly glutamatergic and cholinergic) systems. In addition, the different types of low-frequency rhythms that occur during slow-wave sleep in the disconnected brain are transformed into faster (beta and gamma) oscillations, which are thought to play a significant role in sustaining the mental activity that characterizes brain-active states.

Similarities and basic differences between waking and REM sleep

Global electrical signs differentiating waking from REM sleep

The global signs that characterize the two brain-active states are EEG rhythms, muscular tone, and eye movements associated with sharp waves in brainstem–thalamocortical systems. The tonic EEG activation associated with fast rhythms (usually 20–60 Hz) in waking is virtually identical to that in REM sleep, and distinguishes both of these states from the low-frequency (less than 15 Hz) oscillations that prevail during slow-wave sleep. The other tonic aspect, muscular atonia, specifically distinguishes REM sleep from the other two major states of vigilance (Jouvet & Michel, 1959) and is taken as the cardinal sign of REM sleep. Finally, phasic eye movements are voluntary during waking and occur as involuntary ocular saccades in REM sleep, when they are accompanied by spiky ponto-geniculo-occipital (PGO) potentials in brainstem and thalamocortical systems.

In neuroscience, the term gating can assume various meanings depending on the level of analysis. At the level of ionic channels, gating refers to the transition between two or more conformational states of channel proteins. At the neuronal or networks level, gating refers to changes in responsiveness and in inhibitory processes during different behavioral states. In both instances, the causative events and their functional consequences can vary widely. This book focuses on gating in the thalamocortical and amygdalocortical systems.

In the thalamocortical system, gating was often used to describe the blockade of signal transmission from the external world to the cerebral cortex during disconnected states, such as slow-wave sleep. In this monograph, we also discuss evidence that, despite absence of information from the external world, the behavioural state of slow-wave sleep is associated with the processing of internally generated signals and with synaptic plasticity. We also argue that these events may lead to the consolidation of memory traces acquired during the waking state as well as to a form of consciousness expressed by dreaming mentation. The opening of thalamic gates during brain-active states of waking and REM sleep changes the excitability of cortical neurons and leads to different forms of mentation.

In the amygdalocortical system, gating refers to how the transmission of sensory inputs is modulated according to their emotional significance. This process leads to alterations not only in behavioural responsiveness, but also in memory consolidation.

In this chapter we discuss the organization, connectivity and neuronal properties of different modulatory systems located in the brainstem core, hypothalamus and basal forebrain. These systems exert widespread effects on the thalamus, neocortex, perirhinal cortices, amygdala and related subsystems. Typically, they have an activating effect on target neurons. Activation is defined as a state of readiness in cerebral networks, a state of membrane polarization which brings neurons closer to firing threshold, thus ensuring reliable synaptic transmission and quick responses (Steriade, 1991), without, however, losing sculpting inhibitory processes of short duration that are necessary during the adaptive state of waking (Jasper, 1958; Steriade, 2003c).

One of the major points in this chapter is the demonstration that none of the ascending activating systems is the awakening ‘centre’. In fact, no wake or sleep state can be said to have a centre and even few, if any, components of waking–sleep states have a centre. Functionally, a neural centre may be thought of as subserving only one function. A behavioural state centre would imply a group of neurons, homogeneous in their input–output organization and chemical code(s), and having the required pathways to control the activity of the final effectors of the events involved in that behavioural state (Steriade & McCarley, 2005). Among additional criteria, the centre, when deafferented from its major inputs, should continue to generate the state or to exhibit some of the defining electrographic signs of the state.

We now discuss how external and internal signals are processed in the thalamus and cerebral cortex during the state of slow-wave sleep and during different types of anaesthesia that mimic this sleep stage. Basically, signals from the periphery are obliterated through synaptic inhibition within the thalamus and cannot be transferred to cortex, whereas corticocortical and corticothalamic circuits remain active despite disconnection from the external world. This dissociation explains why, in the absence of information from the external world, the behavioural state of slow-wave sleep is associated with processing of internally generated signals and even with synaptic plasticity. These processes may lead to consolidation of memory traces acquired during the wakeful state as well as to a form of consciousness expressed by dreaming mentation. However, prior to discussing these topics, this chapter will first describe the spontaneously occurring brain oscillations and neuronal firing patterns that characterize slow-wave sleep.

Brain oscillations during slow-wave sleep and anaesthesia in animals and humans

The three major types of brain rhythms, which appear in the state of slow-wave sleep in experimental animals and humans, are mainly generated as a consequence of the reduction in firing rates of brainstem reticular and basal forebrain activating neurons that project to the thalamus and cerebral cortex (see Chapter 5). This relation was observed by recording cellular activity in the upper reticular core and mesopontine cholinergic nuclei during the transition from wakefulness to slow-wave sleep.

This chapter describes the structure and connections of the rhinal cortices and medial prefrontal cortex. We focus on this particular subset of cortical areas because they play a key role in the formation of declarative memories. Indeed, the rhinal cortices are the gateway to and from the hippocampal formation. However, the available evidence suggests that they are not simple relays but instead filter or select inputs. Although the computational rules underlying this function still elude us, it is clear that this process is altered in emotionally arousing conditions. Indeed, memory formation for emotional charged material is generally improved and much evidence suggest that inputs from the amygdala and medial prefrontal cortex mediate this facilitation of memory by emotions. This chapter summarizes data about the structure, connectivity, and physiological properties of these cortical regions. How they interact in memory formation will be considered in Chapter 9.

Cytoarchitectural organization and cell types

Rhinal cortices

The rhinal cortices occupy a strategic location in the temporal lobe because they relay most sensory inputs from the neocortex to the hippocampus. Moreover, the rhinal cortices represent the main return path for hippocampal efferents to the neocortex (reviewed in Witter et al., 2000). Although their precise contribution, compared with that of the hippocampus, remains debated (Brown & Aggleton, 2001), it is clear that the rhinal cortices are not simple relay stations. Indeed, rhinal neurons exhibit patterns of memory-related activity distinct from those seen in the hippocampus (reviewed in Suzuki, 1996; Eichenbaum, 2002).

To discuss gating processes in the thalamus during different normal and pathological conditions (see Chapters 6 and 7), we should first describe the types of neurons and neuronal networks as well as the modulation of intrinsic properties of thalamic neurons by synaptic activities in various behavioural states.

Nuclear systematization, morphology and immunoreactivity of thalamic cells

Thalamic nuclei can be systematized into sensorimotor (or relay), association, intralaminar, and reticular neuronal aggregates. The term relay indicates that those nuclei, among them visual lateral geniculate (LG), auditory medial geniculate (MG), and somatosensory ventroposterior (VP), transfer to cerebral cortex specific sensory signals arising in the ascending afferent pathway. This does not imply that such nuclei operate as mere relays, as if nothing would change between activities in afferent fibres and in thalamocortical axons. Indeed, the presence of local-circuit inhibitory neurons in various nuclei and the relations that thalamic relay neurons entertain with thalamic reticular (RE) inhibitory neurons, account for integrative processes in thalamic relay nuclei, mainly consisting of response selectivity higher than that recorded at prethalamic levels.

Before discussing the morphology, connections and properties of different neuronal classes in the thalamus, a brief account of the major thalamic nuclei is necessary. Figure 1.1 illustrates the nuclear groups in the cat, a species of choice for the study of many topics discussed in this monograph.

In this chapter, we discuss mental disorders that arise from alterations in brainstem–thalamic, thalamocortical and amygdalocortical neuronal circuits. Some of these abnormal activities result from diminution or pathological increase in the number of neurons, which is reflected in their functional output and leads to substantial changes at the level of their targets.

Alterations in brainstem-thalamic and thalamocortical neuronal circuits, with emphasis on hallucinations and schizophrenia

We will first describe what is known about the role of thalamic morphological and functional modifications in the generation of autism, sleep disorders, and loss of consciousness in absence epilepsy, and will next focus on the brainstem–thalamic and basal ganglia circuits that are hypothesized to generate hallucinations in schizophrenia.

The expression of autistic disorders was ascribed by some authors to abnormal activities in some cortical areas and the thalamus. The highly attenuated dreaming in patients with autism and Asperger's syndrome (a developmental disorder regarded as equivalent to autism) was related to the ventromedial frontal cortex, which, when lesioned, gives rise to diminished dreaming. The thalamus was also implicated in autism as anatomical changes and functional alterations through this anteroom to the cerebral cortex may involve disturbances of attention and sensory gating. Thus, in autism, the thalamic volume was found to be significantly different (mainly reduced) relative to normal control subjects (Herbert et al., 2003; Tsatsanis et al., 2003); and emotion processing results in lower regional cerebral blood flow in some cortical areas and in the thalamus (Hall et al., 2003).

The significance of neuronal oscillations in the amygdala and related cortices

Even when deprived of sensory stimulation, neurons hum continuously. That is, their membrane potential fluctuates constantly and the depolarizing phase of these oscillations sometimes gives rise to action potentials. Interestingly, this spontaneous activity is not random. Correlated neuronal events, occurring in a pulsatile or oscillatory manner, can be measured in the extracellular space as currents. These events result from non-linear interactions between the intrinsic membrane properties of neurons and the particular properties of the synaptic network of which they are a part (Llinás, 1988). Oscillations in various frequency ranges are observed in different brain regions; these rhythms vary depending on the behavioural state (Buzsáki et al., 1983; Steriade, 1997a).

The relevance of oscillations to brain function stems from the fact that neuronal events underlying cognition are embedded in these endogenous rhythms. Stated otherwise, one cannot disentangle oscillations from coding in large neuronal ensembles. Moreover, during sleep, when the brain is largely disconnected from the outside world, neurons generate a variety of oscillations and synchronized population bursts that are thought to play a critical role in memory consolidation (see Section 6.3). Finally, because related parts of the brain tend to display similar oscillations, the analysis of spontaneous oscillatory activity can reveal functional kinship among brain structures.

This chapter compares the neuronal oscillations displayed by the amygdala and related cortices.

This list of terms is to start you off. A much fuller glossary of terms is given in Sprinthall (2003).

Degrees of freedom (d.f) gives the number of scores that are free to vary once certain restrictions have been placed on the data. To illustrate the point, when placing eggs into a standard egg box that takes six eggs (n), only one depression would be left for the sixth egg, allowing no ‘freedom’ about where it can be placed. Hence the degrees of freedom are (n − 1) = 6 − 1 = 5. The larger is the sample size, the greater is the number of degrees of freedom. In statistical analysis a degree of freedom is lost for every additional treatment that is added.

Effect size refers to the magnitude of an effect (or, in medical parlance, the ‘clinical significance’). Effect size and statistical significance are quite separate matters and the level of statistical significance does not, as is often supposed, directly measure the magnitude or scientific importance of the observed result. A correlation or a difference can be very small in size yet highly statistically significant, provided the sample size is large enough (see Chapter 11 for further discussion).

Error: Measurement error is the combined error that results from inevitable imperfections and variability in the process of measurement.