Brain sections were processed for in situ hybridization following a variation of the standard procedure using digoxigenin-labeled riboprobes (Medina et al., 2004; García-López et al., 2008; Abellán and Medina, 2009). The riboprobes were synthesized from cDNAs of different mouse or chicken genes.

The cDNAs from mouse genes were obtained from other laboratories:

The cDNAs from chicken genes were purchased [cDNA ESTs purchased from ARK-genomics (Roslin Institute; Midlothian, UK) or Geneservice Limited (Cambridge, UK)], or obtained from other laboratories, as indicated below. The purchased clones were obtained from the BBSRC ChickEST Database (Boardman et al., 2002), and have a corresponding Genbank accesssion number:

We used PCR to obtain the DNA template employed for synthesizing the riboprobe. We synthesized the antisense digoxigenin-labeled riboprobes using Roche Diagnostics's (Mannheim, Germany) protocols for the genes mentioned above. Before hybridization, the sections were abundantly washed in PBS containing 0.1% Tween-20 (PBT 1X), prehybridized in hybridization buffer (HB) for 2 h at 58°C, and then hybridized in HB containing the riboprobe overnight at 58°C (0.5–1 μg/ml, depending on the probe and embryo size). The hybridization buffer contained 50% of deionized formamide, 1.3X standard saline citrate (SSC; pH 5), 5 mM ethylene-diamine-tetraacetic acid (EDTA; pH 8.0; Sigma-Aldrich, Steinheim, Germany), 1 mg/ml of yeast tRNA (Sigma-Aldrich), 0.2% Tween-20, 100 μg/ml of heparin (Sigma-Aldrich), completed with water (free of RNAase and DNAase; Sigma-Aldrich). Following hybridization, the sections were washed with a mix 1:1 of MABT 1X (1.2% maleic acid, 0.8% NaOH, 0.84% NaCl, and 0.1% Tween-20) and HB at 58°C during 20 min and washed abundantly at room temperature with MABT 1X (about 2 h). Following this, the sections were blocked with a solution containing blocking reagent (Roche), MABT 1X and sheep serum (Sigma) for 4 h at room temperature, then incubated in an antibody against digoxigenin (alkaline-phosphatase coupled anti-digoxigenin; diluted 1:3500; Roche Diagnostics) overnight at 4°C, later washed with MABT 1X and finally revealed with BM purple (Roche Diagnostics). Sections were then mounted on glycerol gelatine (Sigma).

Some series of chicken embryonic brain sections (E8–E12) were processed for immunohistochemistry, following a procedure previously described (Abellán and Medina, 2009).

In order to detect radial glial fibers in chicken, we used a monoclonal antibody against chicken vimentin (H5 from Developmental Hybridoma Bank, Iowa, USA; Herman et al., 1993). The specificity of this antibody has beed shown by the manufacturer using Western blot (labeling a band of roughly 52 kDa, corresponding to the protein vimentin).

The immunohistochemical procedure was as follows. After washing in PBS, the sections were incubated in the primary antibody, diluted 1:50 in PBS containing 0.3% Triton X-100, for 2 days at 4°C, under constant and gentle agitation. Then, the sections were washed and incubated in a secondary antiserum for 1 h at room temperature (biotinylated goat anti-mouse IgG; diluted 1:200; Vector, Burlingame, CA, USA). Following this, the sections were washed and incubated in the avidin-biotin complex (ABC kit; Vector; 0.003% dilution) for 1 h at room temperature. Finally, the immunolabeling was revealed by 0.05% diaminobenzidine (DAB; Sigma-Aldrich, Steinheim, Germany) in 0.05 M Tris buffer (pH 7.6), containing 0.03% H₂O₂.

Digital photographs were taken on a Leica microscope (DMR HC) equipped with a Zeiss Axiovision digital camera. Digital images were adjusted for brightness/contrast using Adobe PhotoShop and figures were mounted and labeled using Macromedia FreeHand 10.

Finally, the nomenclature used in the present study for the chicken telencephalon generally followed that proposed by Reiner et al. (2004), except for developmental units, hippocampal subdivisons, and the entorhinal cortex, for which it followed Redies et al. (2001), Puelles et al. (2007), Abellán et al. (2009). For the mouse embryonic brain we primarily followed Jacobowitz and Abbott (1997), and for the mature mouse hippocampal complex, we followed Paxinos and Franklin (2004) and Witter (2012).

The mRNA expression patterns of all genes analyzed were largely conserved between mouse and chicken, although some differences were also appreciated. During very early development (E11.5 in mouse; E6-E7 in chicken), Lef1 was strongly and distinctly expressed in the ventricular zone (vz) of the medial pallium (MP) and the pallial septum (PSe) (chicken: Figures 6A,B), although it also showed generally weak expression in the vz of other pallial sectors. By E13.5 in mouse (Figures 1A,B,D,E) and E8 in chicken (Figures 6C, 7A), Lef1 became primarily restricted to the vz of the medial pallium and pallial septum (Tables 1A, 2A).

At E13.5–E15.5 in mouse (Figures 1–3) and E8-E9 in chicken (Figures 6, 7), the medial pallial sector was characterized by strong or moderate vz expression of Lef1 (Figures 1, 3, 6C, 7A), Lhx2 (Figures 2A–D, 6D,E), and Lhx9 (Figures 2E–K, 6F, 7B,C), but not Lhx5 (Figures 1C, 6G). This feature allowed the distinction of the medial pallium from other progenitor pallial sectors, such as: the pallial septum (PSe), expressing Lef1 (Figures 1A, 6B), Lhx2 (Figure 6E), Lhx9 (Figure 6F), and also Lhx5 (Figures 1C, 6G) in the vz; the dorsal pallium (DP), expressing strongly Lhx2 in the vz (Figures 2A–D, 6D), but not Lef1 (except its medialmost, cingulate-related area; Figures 1A,B,D), Lhx9 (Figures 2E–G) or Lhx5 (Figure 1C) (data on Lhx2, Lhx9, and Lhx5 in chicken DP is published in Abellán et al., 2009, 2010); and the lateral (LP) and ventral (VP) pallia, showing generally weak expression of Lhx2 in the vz (Figures 2A, 6D,E), but no vz expression of Lef1 (Figures 1A,B, 6C) nor Lhx5 (Figures 1C, 6G) (summarized in Tables 1, 2; for the lateral pallial sector, we followed a recent redefinition done by Puelles, 2014).

Based on the combinatorial gene expression patterns studied here and on published data (Puelles et al., 2000, 2007; Medina et al., 2004; Abellán et al., 2009; Puelles, 2014), we tentatively distinguished two new pallial sectors, which we named the dorsolateral caudal pallium (DLP) and the ventrolateral caudal pallium (VLP) (Tables 1, 2). The DLP was previously described in chicken as a distinct subdivision belonging to either the lateral pallium (Puelles et al., 2007) or ventral pallium (Abellán et al., 2009), and has been called temporo-parieto-occipital area or pallium externum in some studies (for example, Veenman et al., 1995; Atoji and Wild, 2005). In contrast to the dorsal and lateral pallia, the DLP expressed Lhx9 in the mantle throughout development (Figures 6F, 10A,H,I) and, for this reason, was previously suggested to be part of the ventral pallium (Abellán et al., 2009). However, in contrast to the ventral pallium, the DLP showed abundant vz/mantle expression of Emx1 (chicken: named CDLx in Figure 10 n of Puelles et al., 2000) and Lhx2 (Figure 10B). In addition, in contrast to the medial pallium, the DLP did not express Lef1 (Figure 6C, Table 2). Herein, we tentatively identified a comparable pallial subdivision in the mouse, giving rise to the lateral entorhinal cortex (LEnt), with similar topological position and genetic profile [no expression of Lef1 in vz. (Figures 1E, 3F,G), but expressing Lhx9 in the mantle (Figure 2H) and Emx1 in vz/mantle (see Allen Developing Mouse Brain Atlas); Table 1]. On the other hand, the VLP [for the moment only identified in chicken, and giving rise to the arcopallium (A)] differed from the ventral pallium [giving rise to the nidopallium (N) and piriform cortex (Pir)] for its strong expression of Lhx9 (Figures 7B,C, 9E, 10D,F,J; Abellán et al., 2009), Lhx2 (Figures 7D, 10E), and Emx1 (see Figure 10p in Puelles et al., 2000) in both the ventricular zone and mantle (Table 2).

In addition to the expression in the vz, most derivatives of the medial pallium of mouse and chicken also showed moderate to strong expression of Lef1 and Lhx2 at least during early and intermediate developmental stages, while some or many of them also expressed Lhx9 (see details in next sections; Tables 1, 2). These features, linked to the molecular identity of the medial pallial vz, helped to identify and compare the medial pallial derivatives between mouse and chicken. The results on the combinatorial expression of Lef1 and other developmental regulatory genes in the developing hippocampal complex (including hippocampal formation and entorhinal cortex) are explained in detail below, first for the mouse (Figures 1–5) and then for the chicken (Figures 6–11).

In this study, we used the combinatorial expression patterns of seven developmental regulatory genes (Lef1, Lhx2, Lhx9, Lhx5, Lmo3, Lmo4, and Prox1), together with analysis of topological position and published data on these or other genes, to identify the medial pallial derivatives and define its major subdivisions in mouse and chicken, and to compare such subdivisions between both species. In both mouse and chicken, the medial pallium is defined as a pallial sector adjacent to the cortical hem (expressing cLhx5 and cWn8b; Figures 6G, 8F) and roof plate/choroid tela (expressing cWnt8b, Figure 8F), showing moderate to strong vz expression of Lef1, Lhx2, and Lhx9, but not Lhx5, at least during early developmental stages. The topological position and the combination of genes at early developmental stages make the medial pallium different from other neighboring pallial sectors, such as the pallial septum (which additionally expresses Lhx5), the dorsal pallium (which does not express Lhx9 in the vz, and is mostly devoid of Lef1 except its medialmost, cingulate/retrosplenial areas), the dorsolateral caudal pallium (which does not express Lef1 nor Lhx9 in the vz), and the ventral/ventrolateral-caudal pallia (which do not express Lef1 in the vz). Based on this gene combination, often also present in the mantle, we propose that the indusium griseum, the hippocampal formation (DG, CA fields, and subiculum), the medial entorhinal cortex, and part of the amygdalo-hippocampal transition area of mouse are medial pallial derivatives. It is likely that the presubiculum and parasubiculum also derive from the medial pallium (see these areas expressing Lhx2 in Figure 5E″), although our data were insufficient to clearly determine expression of Lef1 in these areas. In the chicken, based on the same position and gene combination, we propose that the hippocampus (including the V-shaped area and the ventral hippocampus), the APH (including its caudolateral part, often called CDL; Reiner et al., 2004; Atoji and Wild, 2005), the entorhinal cortex, and the amygdalo-hippocampal transition area are medial pallial derivatives.

The genes Lhx2, Lhx9, and Lef1 were previously described to be expressed in the developing hippocampal formation of mouse (Lhx2: Porter et al., 1997; Bulchand et al., 2001; Monuki et al., 2001; Vyas et al., 2003; Lhx9: Rétaux et al., 1999; Vyas et al., 2003; Abellán et al., 2009; Lef1: Galceran et al., 2000; Choe et al., 2013). Herein, we provide more details on their expression in other medial pallial derivatives, such as the indusium griseum, part of the amygdalo-hippocampal transition area and the medial entorhinal cortex. The common origin with other parts of the hippocampal formation may explain some of their similar features and connections (see discussion for the entorhinal cortex below).

The present study is the first one that uses the three genes in combination, in a comparative context and in a comprehensive way, for trying to identify the medial pallial derivatives in the chicken. There are previous, separate reports of expression of these genes in the developing medial pallium of chicken, but none of these showed enough detail (cLhx2 and cLhx9: Abellán et al., 2009; this study was centered in the ventral pallium; see also data of Lhx9 in the zebra finch: Chen et al., 2013) and/or signal quality (cLef1: Gupta et al., 2012). Based on the combinatorial expression patterns presented here, the chicken medial pallium is larger than previously thought since it includes not only the hippocampus (including the V-shaped area and the ventral hippocampus) and medial parts of APH (our APHm, APHi, APHl; simply named APH in the proposal of the Avian Brain Nomenclature Forum; Reiner et al., 2004), but also the caudolateral part of APH, the entorhinal cortex, and the amygdalo-hippocampal transition area. The caudolateral APH (APHcl, using the nomenclature of Redies et al., 2001; Puelles et al., 2007) is called the dorsolateral corticoid area by many authors (CDL; Reiner et al., 2004; Atoji and Wild, 2005). Its medial pallial origin possibly explains its three-layered cytoarchitecture similar to other APH areas (Redies et al., 2001; Puelles et al., 2007), and its extensive connections with other parts of the APH, as shown in pigeons (Atoji and Wild, 2005). Based on its connections, Atoji and Wild (2005) proposed that the APHcl/CDL is comparable to the cingulate cortex of mammals, which primordium also expresses Lef1 during development (present data). However, while the APHcl/CDL derives from the medial pallium (having vz expression of Lef1, Lhx2, and Lhx9), the cingulate cortex and other parts of the neocortex derive from the dorsal pallium (showing lack of expression of Lhx9 at the vz), which disfavors the homology of these two structures. Our data support that APHcl/CDL is really a medial pallial derivative and, as such, part of the avian hippocampal complex; therefore, we recommend to call it simply APHcl and to abandon the term CDL, which is confusing because it is also employed by some authors to refer to the DLP (see, for example, Puelles et al., 2007; Belgard et al., 2013). In addition, we found a novel cell group, the APHr (maybe comparable to the apical APH of Puelles et al., 2007), which could be distinguished by its intense expression of Lef1 from E10 onwards. While Lef1 started to be downregulated in most of the medial pallium, its expression was intensified in APHr during intermediate and late embryonic development, and was still seen defining this cell group after hatching (P2). As discussed later, an ectopic migrated part of APHr appears to reach intermediate and caudal hippocampal formation levels, where it appears to correspond to the so-called parvocellular region of Atoji and Wild (2005).

The present data agree with previous claims of homology of the avian hippocampus and APH with the hippocampal formation of mammals (reviews in Dubbeldam, 1998; Reiner et al., 2004; Striedter, 2005; Atoji and Wild, 2006; Papp et al., 2007), which were based on identical topological position and embryological origin (Ariens-Kapper et al., 1936; Källén, 1962; Redies et al., 2001; Puelles et al., 2007), and some similarities in cyto- and chemo-architecture (Erichsen et al., 1991; Montagnese et al., 1996; Tömböl et al., 2000; see also Herold et al., 2014), connections (Benowitz and Karten, 1976; Casini et al., 1986; Székely and Krebs, 1996; Székely, 1999; Atoji et al., 2002; Atoji and Wild, 2004), and function (Bingman et al., 1984, 2003; Sherry et al., 1992; Colombo and Broadbent, 2000; Clayton et al., 2003; Mayer et al., 2013). Similarly to that of mammals, the avian hippocampal formation is involved in episodic and spatial memory, and contains location-specific and other types of cells involved in spatial navigation (Clayton et al., 2003; Bingman and Sharp, 2006). It also shows oscillatory activity similar to the theta rhythms (Siegel et al., 2000), LTP and LTD synaptic plasticity involved in learning and memory, synaptic modification after training, and evidence of adult neurogenesis (reviewed by Papp et al., 2007). Crucial for the argument of homology is that the hippocampal formation has also been identified in reptiles, and was likely present in the common ancestor of amniotes (reviewed by Rodríguez et al., 2002; Papp et al., 2007; Medina and Abellán, 2009).

Our data also agree with more recent proposals of homology based on massive gene expression data in the adult pallium, which show a striking similarity of mouse and chicken medial pallial derivatives regarding their gene expression profile (for example, Belgard et al., 2013). However, some recent studies have revealed that the hippocampal formation of different avian species also shows a genetic expression profile similar to that of the arcopallium during development (Chen et al., 2013) and in the adult (Jarvis et al., 2013). This particularly refers to the expression of the transcription factors Lhx9 and ER81. However, Lef1, which in mammals has been shown to be essential for the development of the hippocampal formation (see above), is expressed in the developing hippocampal formation of chicken (E8–E9), but not in the arcopallium. After E10, Lef1 also starts to be expressed in parts of the mantle of the chicken dorsal pallium (hyperpallium) and lateral pallium (mesopallium). It is important to remember that most developmental regulatory genes are expressed in more than one region; for example, this is so for Emx1, Emx2, Pax6, Lhx2, Lhx9, and ER81, expressed in several pallial subdivisions, but some also in the subpallium and outside the telencephalon, and even outside the nervous system (Puelles et al., 2000; Abellán and Medina, 2009; Abellán et al., 2009; Tzchori et al., 2009; Chen et al., 2013; present data). Their function is region, time, and context dependent. This also applies to Lef1, which is expressed in the brain and other tissues, in complex patterns that change throughout embryonic development and postnatally (present data; Oosterwegel et al., 1993; Nagalski et al., 2013), having roles that are context-dependent (Mao and Byers, 2011). The context relies on the molecular networks present in the tissue, which change between regions and with time. The molecular network present in the tissue at any time is essential for understanding both the interactions between transcription factors or other regulatory proteins and their region- and time-specific function. For this reason, we pay special attention to both the topological position (Nieuwenhuys, 1998, 2009; Striedter, 2005) and the combinatorial expression patterns of regulatory genes seen during early development (see also discussion in Puelles and Medina, 2002; Puelles and Ferran, 2012; Medina et al., 2013). Studies using knockout mice have shown that Lef1 is one of the key actors involved at early stages in the development of the hippocampal formation, but this transcription factor acts in combination with other regulatory proteins, such as Wnt and BMP proteins, produced at the cortical hem and/or roof plate (Galceran et al., 2000; Choe et al., 2013). How Lef1 interacts with Lhx2 (also essential for hippocampal development; Bulchand et al., 2001; Vyas et al., 2003), Lhx9, and ER81 during medial pallial development is unknown. The role of Lef1 in the development of other brain regions outside the medial pallium (such as the hyperpallium/dorsal pallium or the thalamus) is also unknown. Due to its far lateroventral position, the arcopallium appears to be out of the effect of roof plate/cortical hem BMP/Wnt signals (if existent, such effect is likely very weak; see also Medina and Abellán, 2009; Aboitiz and Zamorano, 2013). A partially different network of transcription factors (without the implication of Lef1) is important for arcopallial development (such as Lhx9, ER81, and other), although the hierarchy, interactions and functions of the different factors within the network are still unknown.

The combinatorial expression of Lef1, Lhx2, Lhx9, Prox1, Lmo4, and Lmo3 was useful for defining some molecular features of the major subdivisions of the hippocampal formation, and for comparative purposes. Below we discuss the evidence suggesting the comparison of specific chicken subdivisions with the mammallian DG/CA3 and the CA1/subiculum (Figure 12).

Data in chicken and in different mammalian species show that the majority of the neurons of the hippocampal formation migrate radially (following radial glial fibers) from the medial pallium neuroepithelium (mammals: Eckenhoff and Rakic, 1984; Rickmann et al., 1987; Altman and Bayer, 1990a,b; Li and Pleasure, 2005; chicken, Gupta et al., 2012). The exception to this rule is the case of the GABAergic interneurons that populate the hippocampal formation, which migrate from the subpallium (Pleasure et al., 2000; Cobos et al., 2001). In addition, in chicken a part of the cells of APHr (the ectopic APHr or APHre) appears to migrate tangentially within the medial pallium to occupy more caudal, dorsomedial, and superficial positions (present results). Based on Lef1 expression and radial glial fiber disposition, the APHr vz appears to be located at very rostral APH levels (maybe corresponding to the apical APH of Puelles et al., 2007), where Lef1 occupies the whole mantle (Figure 8C; Figure S2A). A band of Lef1-expressing cells (the APHre) appears to extend from this origin, and progressively occupies more superficial, dorsolateral and caudal positions. At intermediate and caudal levels, the Lef1 expression domain related to APHre lies at the surface of APHm (Figures 8D, 9J, 11G) and appears to correspond to the so-called parvocellular region of the hippocampal formation (Atoji and Wild, 2005). Thus, this observation suggests that the neurons of the parvocellular region arrive at their final destination by tangential migration. Supporting this proposal, this region is avoided by radial glial fibers that produce the underlying APHm (Figure 8I). The relation of APHr/APHre to other hippocampal subdivisions of chicken or other amniotes and the function of this cell group remain unknown.

In mammals, the DG granule neurons follow a special type of radial migration due to deformation of the radial glial fibers at the medialmost pallial edge, during the pallial growth that occurs in later developmental stages (Eckenhoff and Rakic, 1984; Rickmann et al., 1987; Li and Pleasure, 2005; note that some authors do not consider this migration to be radial: Altman and Bayer, 1990a). Such deformation of the radial glial fibers is not visible in the medial pallium of chicken (Gupta et al., 2012; present results of radial glial fibers), possibly because it does not grow as much as in mammals.

Our data in mouse suggest that the two major divisions described in the entorhinal cortex of different mammals originate in separate pallial domains, the MEnt (caudomedially located) from the medial pallium, and the LEnt (rostrolaterally located) from the dorsolateral caudal pallium (Figure 12). In particular, based on the combinatorial expression of Lef1, Lhx2, Lhx9, and Lmo4, the MEnt appears to derive from the same embryonic domain that produces the hippocampal formation. This may explain some of the distinct features found in MEnt (but not LEnt) (Sewards and Sewards, 2003), such as the presence of cells involved in processing spatial cues (grid cells, head-direction cells, and border cells, which respond to specific position, direction and orientation, and are able to precisely map the spatial environment), and its implication in transmitting information on the spatial context of an experience to the hippocampal formation (Leutgeb et al., 2005; Knierim et al., 2013; Zhang et al., 2013). Moreover, lesion experiments have shown that MEnt (but not LEnt) is involved in spatial learning (Sewards and Sewards, 2003). In contrast, LEnt transmits non-spatial information to the hippocampal formation, related to the content of an experience, and is involved in non-spatial learning and memory retrieval (Knierim et al., 2013; Stouffer and Klein, 2013; Tanninen et al., 2013). While the MEnt receives input from the CA1, subiculum, presubiculum/parasubiculum (all of which also contain place or grid cells; Boccara et al., 2010), and from visual neocortical areas related to the dorsal visual stream (the “where” pathway) involved in processing spatial visual information on object location (Wang et al., 2011), the LEnt receives input from visual areas of the temporo-occipital neocortex and/or perirhinal cortex (including area 35) related to the ventral visual stream (the “what” pathway), involved in object identification and recognition (Sewards and Sewards, 2003; Canto et al., 2008; Wang et al., 2011). In both rodents and cats, both the LEnt and MEnt receive direct olfactory bulb input, although the LEnt is the preferential target (Room et al., 1984; Witter and Amaral, 2004). In rodents, LEnt shows important reciprocal connections with the pallial amygdala and possibly provides the amygdala with complex “contextual” information relevant for behavior (McDonald and Mascagni, 1997), but the involvement of MEnt in such connections is very modest (Sewards and Sewards, 2003). Both MEnt and LEnt receive weak auditory input from the temporal neocortex, and are reciprocally connected with areas of the cingulate, retrosplenial, and frontal neocortex (Sewards and Sewards, 2003; Witter and Amaral, 2004).

Are these two entorhinal cortex divisions present in birds? Current data suggest that the so-called entorhinal cortex of birds (Puelles et al., 2007; Abellán et al., 2009; present work) may be comparable to mammalian MEnt (Figure 12). This cortical area, located laterally to the APHcl/CDL, receives olfactory input (Reiner and Karten, 1985; Atoji and Wild, 2014), and has often been considered a caudal continuation of the piriform cortex (for example, Atoji and Wild, 2014). However, in both mouse and chicken, the piriform cortex shows a genetic profile different from that of this avian cortical field: the piriform cortex is characterized by strong expression of Lmo3, Lmo4, and Cdh10, very weak expression of Lhx9, and no expression of Lhx2 and Lef1; in contrast, the avian entorhinal cortex shows moderate to strong expression of Lmo4, Lhx9, Lhx2, and Lef1, while its cortical plate is nearly free of Lmo3 and Cdh10 (Vyas et al., 2003; Abellán et al., 2009; present data). Also, while the piriform cortex is at the surface of the nidopallium and derives from the ventral pallium (Puelles et al., 2007), the so-called avian entorhinal cortex is adjacent to the APH and lateral horn of the lateral ventricle, and appears to derive from the medial pallium (based on position and expression of Lef1 during early development). Based on its embryonic origin, the so-called avian entorhinal cortex may be comparable to mammalian MEnt. As noted above, MEnt also receives a minor direct input from the olfactory bulb at least in some mammals (Room et al., 1984; Witter and Amaral, 2004).

In addition, the avian APHcl/CDL may also be comparable to mammalian MEnt (Figure 12). In pigeon, the APHcl/CDL is reciprocally and extensively connected with the various areas of the hippocampal formation, i.e., DG/CA3 area, APHm, APHi, and APHl (V-shaped area, DM and DL in Atoji and Wild, 2005). Moreover, large lesions involving the CDL produce visuospatial deficits suggesting a similar role to that of mammalian MEnt, although this needs confirmation by smaller lesions or specific inactivation of APHcl/CDL (discussion in Atoji and Wild, 2005). The observed deficits are consistent with the inputs to APHcl/CDL from the visual hyperpallium (Figure 5 of Atoji and Wild, 2005), which has been involved in the “where” analysis of the information (Watanabe et al., 2011). Moreover, the lateral part of APHcl/CDL also receives direct input from the olfactory bulb (Reiner and Karten, 1985; Atoji and Wild, 2014), and for this reason it has been compared to the entorhinal cortex of mammals (Redies et al., 2001; Suárez et al., 2006). Curiously, both the avian APHcl/CDL and the mammalian MEnt include cell aggregates or patches showing neurochemical features different from the surrounding area (birds: Redies et al., 2001; Kovjanic and Redies, 2003; Suárez et al., 2006; mammals: Witter and Amaral, 2004). These patches were also evident in our chicken material at P0 as areas of the cortical plate free of cLmo3 expression. They appear to be formed by cells having the same embryonic birth date and expressing the same types of cell adhesion-mediating cadherins (Redies et al., 2001; Kovjanic and Redies, 2003; discussed by Suárez et al., 2006). The connections and functional significance of these patches remain unknown.

On the other hand, it is uncertain whether the avian field called dorsolateral caudal pallium (DLP) is or is not comparable to the dorsolateral caudal pallial field that produces LEnt in mammals, even if they occupy similar topological positions and share some similar molecular features (for example, in general weak or moderate expression of Lhx2, Lhx9, Lmo3, and Lmo4; present data; Figure 12). The avian DLP is relatively large, and has a cortical-like area at its surface that extends ventrally (Puelles et al., 2007; called caudodorsolateral pallium or CDL by these authors), but apparently does not receive any direct olfactory input (Atoji and Wild, 2014; called temporo-occipito-mesencephalic area or TPO by these authors). This general field receives different types of information from several pallial areas, including the entopallial belt (involved in the what analysis of the information, Watanabe et al., 2011) and the mesopallium (Figure 9 of Atoji and Wild, 2005), and it projects to the avian pallial amygdala (in particular, the caudolateral nidopallium and the arcopallium) and the basal ganglia (Veenman et al., 1995; Kröner and Güntürkün, 1999). It is also connected reciprocally with the hippocampal formation, but less so than the APHcl/CDL (Atoji and Wild, 2005). It would be interesting to investigate whether there is a structure comparable to avian DLP in reptiles, which would contribute to understand its homology across amniotes. For this purpose, it is necessary to have a molecular marker (or a clear combination of them) specific of this pallial sector.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.