The striatum and its position in the basal ganglia loop
The striatum is a large nucleus of the forebrain and the main input station of basal ganglia. Basal ganglia are a set of nuclei located at the interface between the cerebral cortex and diencephalic and midbrain structures. These nuclei are heavily interconnected and form a recursive circuit that is often referred to as a “loop”. Basal ganglia nuclei include the striatum, the globus pallidus, the nucleus accumbens as well as parts of the amygdala. They are intensely connected with other cerebral structures, including the cerebral cortex, the thalamus, and midbrain dopaminergic nuclei, forming together major brain networks involved in motor control, habit formation, reward and addiction. In brief, in response to cortical and thalamic activation, and modulated by dopamine, the striatum triggers two descending pathways that target different basal ganglia nuclei, thereby controlling the selection of appropriate motor sequences. Before describing in more details such complex circuits (see Part II), it is essential to overview the striatum connectivity and its cellular organization.
The striatum is at the crossroad of forebrain connectivity
Main inputs to the striatum
A main input to the striatum is formed by collaterals of descending cortical axons that originate from the entire neocortex (Bunner and Rebec, 2016; Haber, 2014a; Shepherd, 2013; Shipp, 2016; Zold et al., 2012), making this structure the main input station of basal ganglia.
The striatum receives projections from many different regions of the brain and represents the main input station of basal ganglia. Glutamatergic input from cerebral cortex and thalamus is represented by cyan arrows; dopaminergic input from midbrain (SNc, substantia nigra pars compacta; SNr, substantia nigra pars reticulata; VTA, ventral tegmental area) in black; serotonergic input from raphe in blue; GABAergic input from globus pallidus (GP) in red. For clarity, cholinergic projections peduncolopontine nucleus (PPN) is omitted. Dotted lines represent regions that are not on the same sagittal plane. Adapted from Gerfen (1992) to cortical axons, that use Glutamate as neurotransmitter and have therefore excitatory effect, the other main input that target the striatum are:
i. glutamatergic projections from the amygdala and the thalamus (Bunner and Rebec, 2016; Giménez-Amaya et al., 2000; Smith et al., 2004, 2014; Yager et al., 2015);
ii. strong dopaminergic innervation from two midbrain nuclei, the ventral tegmental area (VTA) and the substantia nigra compacta (SNc) (Baydyuk et al., 2013; Vandenheuvel and Pasterkamp, 2008); dopamine has a crucial modulatory role in striatal circuits (see Part II). In addition, the VTA send projections that use γ-Amynobutiric acid (GABA) as neurotransmitter to the ventral part of the striatum (Brown et al., 2012; Van Bockstaele and Pickel, 1995)
iii. cholinergic input from laterodorsal tegmental area and peducolopontine nuclei (Dautan et al., 2014; Hallanger and Wainer, 1988; Silberberg and Bolam, 2015)
iv. moderate serotonergic input from the dorsal raphe nucleus (Mathur et al., 2011)
v. GABAergic projections from arkypallidal neurons of the Globus Pallidus (Dodson et al., 2015; Nóbrega-Pereira et al., 2010).
Striatal output connectivity
A vast majority of striatal neurons use the inhibitory neurotransmitter GABA, making the striatal output entirely inhibitory. These projections target principally two nuclei that are part of the basal ganglia loop:
(i) a nucleus of the midbrain called substantia nigra reticulata (SNr); this structure is connected to many subcortical motor centres and its neurons provide tonic inhibition, preventing motor activation unless their control is blocked
(ii) another nucleus of the basal ganglia, the globus pallidus (GP). It has a “pacemaker” role on the SNr by keeping under tonic inhibition another nucleus, the subthalamic nucleus (STN), which provides strong excitatory signals to SNr neurons.
Therefore, the modulation of SNr activity is the common output of striatal projections. Projections that directly target the SNr form the so-called direct pathway, while projections that terminate on the GP form the indirect pathway. In addition to this basic dichotomy, direct pathway axons send collaterals to the GP, indicating that direct pathway neurons modulate indirect pathway activity and that, as a consequence, the two pathways are not entirely segregated (Cazorla et al., 2014, 2015).
Another pathway, the hyperdirect pathway, connects monosynaptically to a different region of the SN, the substantia nigra compacta (SNc). Only a small subset of striatal neurons, discussed in Part 4.2, form this pathway, which might however possess pathological relevance (Graybiel and Grafton, 2015; Petersen et al., 2016).
Axonal pathways traverse the striatum
The striatum owes its name to the many axonal tracts that cross it. Indeed, this structure is located at the interface between the superficially located cerebral cortex and structures of the subcortical forebrain. In particular, thalamo-cortical and cortico-thalamic projections form a large tract of white matter, called internal capsule, that traverse the striatum. In primates and humans, the internal capsule segregates anatomically the striatum in two structures, the caudate nucleus and the putamen.
Sagittal schematic view of striatal output projections. Striatal output is formed by two parallel GABAergic pathways that target the substantia nigra pars reticulata (SNr) either directly (direct pathway, red arrow) or indirectly by projecting the globus pallidus (indirect pathway, blue arrow). In turn, the globus pallidus (GP) sends inhibitory projections to the subthalamic nucleus (STN) that tonically activates SNr neurons (see text). Direct pathway neurons send collaterals to the GP (short arrows, (Cazorla et al., 2014). The hyperdirect pathway is omitted for clarity.
Striatal projection neurons form direct and indirect pathways
Due to its nuclear structure, the striatum lacks the recognizable organization of layered structures such as the cerebral cortex; moreover, compared to other forebrain structures it possesses relatively few cellular types. In spite of this deceptively simple appearance, however, the striatum is able to integrate a large amount of different input and transform it in two parallel pathways; to understand how this structure is able to control the diverse functions that will be described in Part II, it is important to characterize first its cytoarchitecture.
The striatum is composed by two subtypes of projection neurons that alone accounts for 90-95% of the total number of striatal neurons (Fox et al., 1971; Hull et al., 1981; Levine et al., 1986; Silberberg and Bolam, 2015), plus four classes of GABAergic interneurons and a population of cholinergic interneurons.
Striatal projection neurons
Striatal projection neurons (SPN) are medium-sized GABAergic projection neurons. Morphologically, they are characterized by a rich dendritic tree densely covered with dendritic spines (Bertran-Gonzalez, 2010). They are essentially silent, and have membrane properties that give them a high threshold for activation. Indeed, they tend to remain in a stabilized, hyperpolarized “down state” except when they receive strong and sustained excitation. In this case, when potassium currents are deactivated, SPN enter an “up state” which lets them fire repeatedly at a low rate (Grillner et al., 2005). SPN dendritic spines are described in detail in Section 2.1.3. In brief, these highly plastic structures are contacted by glutamatergic terminals of both cortical and thalamic axons, which contact the head of SPN spines, as well as by dopaminergic terminals, that preferentially contact the spine neck (Day et al., 2008).
Direct and indirect pathway projection neurons have different properties
Despite a similar morphological appearance, two main subtypes of SPN are distinguished by several aspects: in the first place, they send projection to either the SNr (with at least a subset of them sending collaterals to the GP, see (Cazorla et al., 2014) or the GP, thus forming the direct (dSPN) or indirect (iSPN) pathway. Furthermore, dSPN express the D1 dopamine receptor (Drd1) and high levels of neuropeptide SubstanceP, while iSPN express the D2 dopamine receptor (Drd2) and opiate peptide Enkephalin (Gerfen, 1992). Roughly the same amount of dSPN and iSPN compose the total SPN population; however, the two subtypes are randomly intermixed in the dorsal part of the striatum (dSTR, Gangarossa et al., 2013a), while in the ventral striatum (vSTR) they show a more complex distribution (Gangarossa et al., 2013b). It has thus been difficult to investigate the different properties of dSPN and iSPN at the cellular scale up until recently. Indeed, the development of subtype-specific reporter mouse lines (Drd2-EGFP, Drd1-EGFP, Drd1-tom, Drd1-cre, Drd2-cre) has allowed a new level of analysis of dSPN and iSPN (Lobo, 2009; Matamales et al., 2009; Thibault et al., 2013). These studies revealed differences in morphology, electrophysiological properties and input connectivity between the two populations. In particular, dSPN are less excitable than iSPN, which possess more primary dendrites thus receiving a greater number of excitatory synaptic contacts (Gertler et al., 2008; Kreitzer and Malenka, 2007; Tritsch and Sabatini, 2012). Moreover, the two subtypes process differently glutamatergic input, as iSPN synapses have higher release probability (Kreitzer and Malenka, 2007).
In terms of reciprocal connectivity, dSPN are more likely to establish unidirectional connections with each other rather than contacting iSPN. On the contrary, unidirectional iSPN connections to dSPN are more common (Taverna et al., 2008).
Finally, a much debated issue is whether dSPN and iSPN receive input from different sources, in terms of both cortical regions and layers. In particular, while it has previously been shown through quantitative electron microscopy that dSPN and iSPN receive similar input from the cerebral cortex and the thalamus (Doig et al., 2010), retrograde tracing using rabies virus indicated regional differences in cortical innervation of the two subtypes. Somatosensory and limbic areas of the cortex (as well as the amygdala) seem to target preferentially dSPN, while Schematic representation of the main difference between direct and indirect SPN in terms of molecular, synaptic and electrophysiological properties, along with a schematic representation of dSPN and iSPN main projections. dSPN project to the substantia nigra (SN) with at least a subset of them also sending collaterals to the globus pallidus (GP). Conversely, ISPN project to the GP therefore affecting indirectly SN activation.
motor areas would send more projections to iSPN. At the same time, thalamic input would not show spatial preference in targeting of dSPN or iSPN (Wall et al., 2013). Moreover, it has been reported that, in the rat matrix compartment, dSPN and iSPN receive projections from different cortical layers, with intra-telencephalic (layer III-V) cortical neurons contacting preferentially dSPN and pyramidal tract (layer Vb-VI) neurons targeting preferentially iSPN (Reiner, 2010; Deng et al., 2015). Perhaps these differences are linked to the specific early response to sensorimotor stimuli observed in dSPN but not iSPN (Sippy et al., 2015); whatever the case, the issue of differential input to SPN subtypes remains still largely to be investigated, especially in light of recent studies that highlight the complex topographic organization of corticostriatal input (Hintiryan et al., 2016).
Interneurons, or local circuit neurons, modulate the activity of nearby projection neurons. Striatal GABAergic interneurons share an inhibitory role on striatal projection neurons; however, they are divided in at least four main classes based on electrophysiology, morphology and/or the expression of markers:
₋ Fast-spiking interneurons that express the calcium-binding protein parvalbumin (PV)
₋ Interneurons that express neuropeptide Y (NPY); this category comprises two subsets, somatostatin-positive low-threshold spiking and plateau potential (LTSP)-NPY interneurons, and neurogliaform (NGF)-NPY interneurons (Ibáñez-Sandoval et al., 2011).
₋ Interneuron that expresses the calcium-binding protein calretinin (CR)
₋ Tyrosine Hydroxylase (TH) expressing interneurons (Ibáñez-Sandoval et al., 2010).
Finally, giant cholinergic interneurons are the biggest cells in the striatum and the main source of acetylcholine (Ach) in this system. Ach has a complex modulatory function in the striatum, but its main function is to facilitate activation of striatal projection neurons (for review, see Benarroch, 2012). Cholinergic interneurons set a regulatory tone through the tonic release of Ach, acting on muscarinic receptors present on both direct and indirect pathway projection neurons as well as on nicotinic receptors located on striatal afferent projections (Ebihara et al., 2013; Gonzales and Smith, 2015; Tozzi et al., 2011; Wang et al., 2006)).
Anatomical subregions of the striatum related to connectivity
Dorsal striatum versus ventral striatum
The striatum is functionally divided in two regions, the dorsal striatum (dSTR) and the ventral striatum (vSTR) or nucleus accumbens (NAcc), which in turns comprises “core” and “shell” compartments. NAcc core functional organization strongly resembles the dSTR, as it contains randomly intermixed Drd1-expressing SPN and Drd2-expressing SPN. However, the connectivity of the two subtypes is substantially different from the segregated direct/indirect pathways observed in the dorsal striatum. NAcc core targets its main output nucleus, the ventral mesencephalon (VM), either by direct projections or indirectly via the ventral pallidum (VP) (Bock et al., 2013; MacAskill et al., 2014; Yawata et al., 2012); while only Drd1-expressing MSN target directly the VM, both Drd1-expressing and Drd2-expressing SPN project to the VP (Kupchik et al., 2015); therefore, there is less segregation in output projections in NAcc core with respect to dSTR.
Organization of the Nacc shell differs even more from the dorsal striatum: the distribution of Drd1-expressing SPN and Drd2-expressing SPN is very inhomogeneous and associates with the presence of specific accumbal subdivisions (Gangarossa et al., 2013b). Moreover, the shell compartment has very diverse output targets that include VTA, hypothalamus, VP and brainstem.A controversial issue concerns the presence of SPN co-expressing both D1 and D2 dopamine receptors in the NAcc compared to the dSTR (see Gangarossa et al., 2013b). There is general consensus that more SPN co-express D1 and D2 receptors in NAcc shell compared to dorsal striatum and NAcc core; however, estimation of co-expressing neurons varies significantly in different studies (from 5% to around 17% in NAcc shell; for reference, see Kupchik et al., 2015; Thibault et al., 2013).
Striosome and matrix compartments in the dorsal striatum
Within the dSTR, there is another level of organization that defines two major functional subdivisions, striosome and matrix compartments. Historically, striosomes have been defined as areas of dense μ-opioid receptor expression and low acetylcholinesterase labelling, while matrix is composed of neurons that contain a 28 kD calcium-binding protein (CaBP) and high number of somatostatin-positive fibers (Gerfen, 1992). Moreover, Substance P, which is expressed in dSPN, is present at higher levels in striosomes (Tajima and Fukuda, 2013) and mediates dopamine transmission differently in the two compartments (Brimblecombe and Cragg, 2015), while opioid peptide Enkephalin, specific of iSPN, has stronger expression in the matrix (Tajima and Fukuda, 2013). Striosomes develop rostro-caudally across the striatum and are surrounded by the larger matrix compartment, in a pattern that is consistent across individuals. It has been estimated that striosomes add up for roughly 15% of striatal surface (Desban et al., 1993; Fujiyama et al., 2015). Both compartments contain intermixed dSPN and iSPN, although the proportion of dSPN is a bit higher in striosomes compared to the matrix; SPN dendrites are confined within the compartmental borders and only interneurons mediate communication between the two compartments. Several lines of evidence indicate that striosomes and matrix differ in their connectivity: in terms of output, it has been proposed for a long time, and recently demonstrated (Friedman et al., 2015; Fujiyama et al., 2011), that dSPN located within striosomes send specific projections to the pars compacta of the substantia nigra (SNc), forming the so-called hyperdirect pathway (Burguière et al., 2015; Friedman et al., 2015).
Schematic representation of striatal striosome and matrix compartments. Both compartments contain intermixed dSPN and iSPN (red and blue dots). However, striosome and matrix neurons differ for time of generation. Moreover, several molecules are expressed at higher levels in striosomes and matrix, respectively (boxes). Finally, at least a subset of dSPN located in striosomes project to the substantia nigra pars compacta (SNc), forming the hyperdirect pathway (see text), while all matrix dSPN project to the substantia nigra pars reticulata (SNr).
Moreover, striosome receive projections from the limbic cortex, especially orbitofrontal cortex and insula, while input to the matrix arise from a wide area of the cortex, including motor cortex, somatosensory area and parietal lobe (Fujiyama et al., 2015). Moreover, thalamic input is thought to be almost three times stronger in matrix compartment and comes mainly from intralaminar nuclei, while striosomes are targeted by midline thalamic nuclei (for references, see Fujiyama et al., 2015). Despite the attention dedicated to the different connectivity of striosomes and matrix, there is surprisingly few direct evidence supporting the idea that neurons within the two compartments have different functions; this is mainly due to difficulties in specific targeting of striosome or matrix. Only recently, optogenetics studies started to elucidate the function of specific cortico-striosomal circuits in decision making (Friedman et al., 2015), paving the way for a better understanding of the function of the two striatal compartments.
Dorso-lateral versus dorso-medial striatum
Aside of dSTR/vSTR and striosome/matrix subdivision, regionalization of the dSTR on a dorso-lateral to ventro-medial axis has been proposed based on the topographic organization of incoming projection, along with the result of behavioural studies on rodents and primates based on reward processing and habit formation. In fact, glutamatergic input targeting the striatum from the cerebral cortex, the thalamus and the amygdala is topographically organized; the dorsolateral striatum (DLS) receives predominantly sensorimotor-related information, while the NAcc receives visceral-related afferents. The striatal areas between these extremes receive higher-order “associative” information. Moreover, the arrangement of SPN projections to dopaminergic nuclei follows the same organization, albeit with an inverted topography; the DLS projects to the ventrolateral SNr, while medial parts of the dorsal striatum (dorsomedial striatum, DMS) target more dorsomedial parts of the nigra (Voorn et al., 2004). This topographic arrangement is thought to play a role in the acquisition of habits based on reward, which will be investigated in more detail in Part II.
This scheme from Voorn et al. (2004) represents topologically organized input to the anterior striatum from the cerebral cortex and the thalamus. The DLS receives somatotopically organized sensorimotor information (green), the vSTR collects viscerolimbic cortical afferents (red and pink), and striatal areas between these extremes receive information from higher associational cortical areas (blue and purple). Amygdalostriatal and hippocampal projections have mediolateral organization. Abbreviations: ac, anterior commissure; ACd, dorsalanterior cingulate cortex; AId, dorsal agranular insular cortex; AIv, ventral agranular insular cortex; CeM, central medial thalamic nucleus; CL, central lateral thalamic nucleus; IL, infralimbic cortex; IMD, intermediodorsal thalamic nucleus; MD, mediodorsal thalamic nucleus; PC, paracentral thalamic nucleus; PFC, prefrontal cortex; PLd, dorsal prelimbic cortex; PLv, ventral prelimbic cortex; PV, paraventricular thalamic nucleus; SMC, sensorimotor cortex.
“spiralling” connectivity from ventral-medial regions in striatum to dopaminergic nuclei, which in turn project back to DMS (see Burton et al., 2015). The DLS/DMS distinction provides evidence that the dSTR, despite its nuclear organization giving it a homogeneous appearance and the absence of clear-cut borders (except for the striosome-matrix distinction), is a topographically organized structure.
Topologic organization of corticostriatal input
A recent study by Hintiryan et al. (2016) has assessed for the first time the precise topology of corticostriatal connections in mice. This study reveals that the dSTR can be divided in 29 different regions, based on the origin of the cortical input they receive; in addition to that, it shows that at rostral-most and caudal-most levels the dSTR receives mainly projections from high-level, “associative” cortical areas, while at commissural rostro-caudal level the dSTR receives concomitant projections from primary somatosensory and motor cortical areas. At this level, the DLS contains a topologically organized representation of mouse body parts, with different regions representing lower and upper limbs, head, etc. In addition, each region receives input from related motor and sensory cortical areas, indicating that striatal somatotopic areas integrate both sensory and motor information. Finally, protein expression patterning studies are starting to reveal potential molecular mechanisms that drive topological organization of dSTR topographic organization: for example, glycoprotein Ten-m3 is responsible for the clustering of thalamostriatal projections in discrete domains within the striatal matrix (Tran et al., 2015), while receptor tyrosine kinase EphA7, expressed in SPN sub-domains inside the striatal matrix, restricts cortical somatosensory projection domains through a repulsive action (Tai and Kromer, 2014; Tai et al., 2013). Investigating the mechanisms that regulate establishment and refinement of striatal topography will have a great impact on our understanding of the functioning of this structure in normal as well as in pathological conditions.
Roles of the striatum in health and disease
The striatum: from motor control to habit learning
Striatal functioning has been studied for a long time due to its importance in motor pathologies, addition to drugs of abuse, and dopamine-related reward circuits. This led to the formation of a “classic model” of dSTR functioning that focuses on alternate activation of direct and indirect pathways in motor control mediated by different effects of dopamine receptor activation on dSPN and iSPN. However, since dSPN and iSPN are intermixed, it has been difficult to dissect the precise functioning of direct and indirect pathways. The implementation of transgenic mouse lines in last years allowed specific targeting of dSPN and iSPN; an increasing amount of evidence is starting to define a more complex system than what previously assumed and to reveal the mechanisms that underlie striatal role in reward-based learning and habit formation.
The classic “go/no go” model of striatal function
The classic model of basal ganglia circuit function is based on the segregation of information processing in dSPN and iSPN. The two populations would form two parallel pathways – a “go” pathway to promote movement initiation and a movement-inhibiting “no go” pathway, respectively. The model states that excitation of a group of dSPN by corticostriatal and/or thalamocortical afferents provides strong inhibition of neurons of the SNr, the output nucleus of basal ganglia. This will turn off the tonic inhibition of target neurons in a motor centre and consequently activate its motor program. For this reason, dSPN are part of the “go” pathway that facilitates movement. iSPN activation, on the other side, would start the movement-inhibitory “no go” pathway. When iSPN discharge, they inhibit the tonic activity of the GP; this leaves the excitatory (glutamatergic) neurons of STN free to promote activation of SNr neurons therefore inhibiting movement initiation (see Section 2.3).
The most prevalent version of this model predicted that dSPN would be more active during movements, while iSPN would be more active during rest, with dopamine modulating differentially the two pathways via D1 and D2 receptors. The simplicity of this model allowed to explain many of the symptoms observed in neurodegenerative diseases, and was initially reinforced by pathway-specific optogenetic studies showing that a robust activation of direct pathway was sufficient to induce movement, while activation of the indirect pathway led to strong movement inhibition (Kravitz et al., 2010).
More recent studies, however, established that dSPN and iSPN are activated at the same time during movement initiation and are inactive when the animal is not moving (Cui et al., 2013), leading to symmetric activation of direct and indirect pathway target structures (Jin et al., 2014). These findings are still consistent with dSPN role in initiating the “go” response (Sippy et al., 2015), but show that the role of direct and indirect pathways is not just to alternate go/no go signals. Instead, the frame that is emerging is that dSPN could be responsible to select the desired motor program while iSPN would in parallel inhibit competing motor programs; co-activation of the two pathways would then allow selection of the appropriate action sequence (Jin and Costa, 2015).
These findings indicate a more complex role for striatal neurons than the classic model originally accounted for; moreover, they provide supporting evidence for a role of SPN in habit learning based on reward. Still, as predicted by the classic model, direct and indirect pathway activity is influenced by fast excitatory and inhibitory inputs as well as by a slower but extremely powerful modulation by dopamine (DA) controlling SPN excitability and long-term plasticity at synaptic terminals.
This scheme adapted from Pappas et al. (2014) shows the basic organization of basal ganglia loop according to the classic model. Shortly, activation of the direct pathway (in red) following cortical and thalamic activity drives direct inhibition of GABAergic neurons of the substantia nigra reticulata (SNr) and, along with direct cortical excitation, activation of motor neurons (“go” signal). Conversely, indirect pathway activation (cyan) favours tonic inhibition of motor neurons via multi-synaptic circuit (“no go” signal, see text). Activation of the direct and indirect pathways is modulated by dopamine release from the substantia nigra compacta (SNc). Many connections have been omitted for clarity.
Table of contents :
1 Anatomy of the striatal mosaic: multiple levels of organization at the core of the brain
1.1 The striatum and its position in the basal ganglia loop
1.2 The striatum is at the crossroad of forebrain connectivity
1.2.1 Main inputs to the striatum
1.2.2 Striatal output connectivity
1.2.3 Axonal pathways traverse the striatum
1.3 Striatal projection neurons form direct and indirect pathways
1.3.1 Striatal projection neurons
1.3.2 Direct and indirect pathway projection neurons have different properties
1.3.3 Striatal interneurons
1.4 Anatomical subregions of the striatum related to connectivity
1.4.1 Dorsal striatum versus ventral striatum
1.4.2 Striosome and matrix compartments in the dorsal striatum
1.4.3 Dorso-lateral versus dorso-medial striatum
1.4.4 Topologic organization of corticostriatal input
2 Roles of the striatum in health and disease
2.1 The striatum: from motor control to habit learning
2.1.1 The classic “go/no go” model of striatal function
2.1.2 The role of dopamine in SPN excitability
2.1.3 Synaptic plasticity and reward in SPN
2.1.4 The role of dorsal striatum in habit learning based on reward
2.2 Pathologies associated with striatal function
2.2.1 Neurodegenerative disorders: Parkinson’s Disease, Huntington’s Disease
2.2.2 Cognitive dysfunctions: schizophrenia, autistic spectrum and obsessivecompulsive disorders
3 Development of the striatum and related structures
3.1 Establishment of progenitor domains in early embryonic development
3.1.1 Regionalization of the forebrain
3.1.2 Proliferative domains within the ventral telencephalon
3.2 Neuronal progenitors in the ventral telencephalon
3.3 Genetic control of neurogenesis in the ventral telencephalon
3.4 Radial and tangential migration in the ventral telencephalon
3.4.1 Radial migration of projection neurons
3.4.2 Tangential migration of interneurons
3.4.3 Tangential migration of projection neurons
3.4.4 Tangential migration from the lateral ganglionic eminence generates guidepost cells for thalamo-cortical axons
3.5 Differentiation of striatal projection neurons
3.5.1 Molecular regulators of striatal projection neuron (SPN) development
3.5.2 Known molecular actors in dSPN development
3.5.3 Specification and Maintenance of iSPN rely on epigenetic control
3.6 Development of striatal compartments and circuitry
3.6.1 Contribution of striatal projection neurons to striosomes or matrix is defined by time of birth
3.6.2 Survival of striatal projection neurons depends on external sources
3.6.3 Dendritogenesis and Synaptogenesis in striatal projection neurons
4 Transcription factor Ebf1 is a master gene for cell fate specification and neuronal differentiation
4.1 General presentation of the Ebf gene family
4.1.1 Ebf genes in mammals
4.1.2 Roles of COE genes in the specification and differentiation of different cell lineages
4.2 Function of Ebf1 in B-cell specification
4.3 An emerging role for Ebf1 in direct pathway neurons development