The perirhinal cortex (PRC) is located along the rhinal sulcus and can be divided anatomically into Brodmann areas 35 and 36 (Brodmann, 1909). It consists of six cortical layers, whereby layer IV of area 36 contains only a few granule cells and area 35 is considered to be an agranular cortex (Burwell et al., 1995; Burwell, 2001; Kealy and Commins, 2011).

The PRC has been reported to support the integration of multisensory information, emotional aspects of learning such as fear conditioning, spatial memory, and also perception (for review: Kealy and Commins, 2011). The role of the PRC in object recognition memory is undeniable: over the last decades many studies, using different behavioral approaches alone or in combination with PRC lesions, single unit recordings, or immediate early gene (IEG) analysis, support its involvement in novelty detection, as well as the encoding, consolidation, and retrieval of object recognition memory (Mumby and Pinel, 1994; Zhu et al., 1995a; Aggleton and Brown, 2005; Albasser et al., 2015).

Sensory information reaches areas 35 and 36 of the PRC in a differential manner, e.g., inputs from polymodal associational areas are, in general, very strong, but they target area 36 more heavily, whereas unimodal associational areas project to a higher degree to area 35 (Burwell and Amaral, 1998a). As an essential part of the parahippocampal formation, the PRC serves as a transition area between neocortical areas and the hippocampus that interfaces indirectly, via the entorhinal cortex, but also directly transmits information to the hippocampus (for review: Kealy and Commins, 2011). PRC projections originate in most anatomical layers and project stronger to the lateral entorhinal cortex (LEC) than to the medial entorhinal cortex. Weak back projections, traveling mostly to the rostral PRC, arise from layers III and V of mostly the LEC (Burwell and Amaral, 1998a, b). Direct projections to the hippocampus are sparse and are presumed to originate only in superficial PRC layers (Furtak et al., 2007; Agster and Burwell, 2013). In contrast, projections from the hippocampus (stronger from the ventral than the dorsal aspect) target deep layers of preferentially area 35 and to a smaller degree area 36 (Agster and Burwell, 2013). These direct connections, together with the indirect connections to the hippocampus lead one to suspect that a layer-specific engagement in object recognition memory may occur in areas 35 and 36. In line with this possibility, a layer-specific distribution of the different types of glutamatergic receptors has been described for neocortical regions, including the PRC, suggesting layer-specific encoding of e.g., object information (Ziakopoulos et al., 1999; Winters and Bussey, 2005; Palomero-Gallagher and Zilles, 2015).

Anatomical and functional differences along the rostro-caudal axis have also been described (Burwell and Amaral, 1998b; Otto et al., 2000; Burwell, 2001; Furtak et al., 2007; Agster and Burwell, 2009; Albasser et al., 2013; Sethumadhavan et al., 2020). Novel object exposure has been reported to trigger a change in immediate early gene (IEG) expression in caudal areas 35 and 36 (Zhu et al., 1995b; Albasser et al., 2010). Object recognition also triggers gene encoding in these PRC areas: Burke and colleagues examined IEG expression in layer V, or neuronal activity in layer II/III and V of the areas 35 and 36 (Burke et al., 2012a, b), and demonstrated that novel and familiar exploration of objects results in enhanced Arc mRNA expression in layer V (Burke et al., 2012a). In our previous work, we reported that layer II of the caudal area 35, along with the postrhinal cortex (POR), engages in the encoding of information about large objects (Sethumadhavan et al., 2020). Given that the PRC strongly projects to the entorhinal cortex that in turn projects to the hippocampus (Burwell and Amaral, 1998b; Agster and Burwell, 2013; Doan et al., 2019), it is not unlikely that object dimensions also result in layer-specific neuronal encoding within the PRC. This aspect is as yet unclarified. However, it has been reported that novel learning about spatial configurations of large and overt (macroscale) or small and discrete (microscale) objects enables both subregion-specific long-term depression (LTD) and nuclear gene encoding in specific neuronal subcompartments of the hippocampus (Manahan-Vaughan and Braunewell, 1999; Kemp and Manahan-Vaughan, 2008; Hagena and Manahan-Vaughan, 2011; Hoang et al., 2018, 2021).

In the current study, we scrutinized the influence of microscale and macroscale item learning on immediate early gene expression in the PRC to specify the contribution of the layers of areas 35 and 36 to these specific forms of learning. To do this, we analyzed nuclear IEG expression in layers II, III, and (I)V-VI of the middle and caudal areas 35 and 36 that was triggered by novel microscale or macroscale item learning. Our results reveal that layers III and V-VI of area 35 engage in the encoding of both micro- and macroscale items, whereas layer III of area 36 is specifically involved in microscale item encoding.

Seven-to-nine-week-old male Wistar rats were used for this study. Animals had ad libitum access to water and food and were housed in temperature (22 ± 2°C) and humidity (55 ± 5%) controlled containers (Scantainer, Scanbur Technology A/S, Karlslunde, Denmark) on a 12 h light/12 h dark cycle (lights on from 7 a.m. to 7 p.m.). All experiments were approved in advance by the animal ethics authority of the Federal government of the state of North Rhine-Westphalia (NRW; Landesamt für Arbeitsschutz, Naturschutz, Umweltschutz und Verbraucherschutz, NRW) and carried out according to the European Communities Council directive of 22 September 2010 (2010/63/EU) for the care of laboratory animals. All efforts were made to reduce the number of rats used.

It is well described that the PRC is involved in the encoding of object recognition memory (Mumby and Pinel, 1994; Zhu et al., 1995a; Aggleton and Brown, 2005; Albasser et al., 2015). Some previous studies examined IEG expression separately for areas 35 and 36 (Albasser et al., 2010, 2013; Burke et al., 2012a). Nevertheless, these studies could not detect differences in the engagement of areas 35 and 36 in novel object learning, although the intrinsic connectivity of both areas, as well as several afferent connections, are distinct (Burwell and Amaral, 1998a, b; Burwell, 2000; Furtak et al., 2007). The present study aimed to examine the effect of novel learning about micro- and macroscale items in the layers (II, III, V-VI/IV-VI) of the middle and caudal areas 35 and 36. We assessed nuclear Arc mRNA expression in the region of interest in the PRC after novel (micro- or macroscale) item exploration.

Here, we verified that in all conditions where IEG expression was subsequently assessed, that adequate and equivalent object exploration occurred (Table 2). Thus, as reported previously (Sethumadhavan et al., 2020) animals paid equal attention to all of the objects placed in the holeboard holes in the microscale paradigm, and to all of the objects in the macroscale paradigm (Table 2A). The animals also spent almost all of the 5 assigned minutes actively exploring in the paradigms, whereby the degree of exploration was equivalent in the two paradigms (Table 2B). The number of rears was higher in the microscale, compared to the macroscale condition (Table 2B) condition. Effects were distinct from the control condition where animals simply remained stationary and at rest in one corner of the experimental chamber. Behavioral data were also reported in Sethumadhavan et al. (2020).

A role for the PRC in item recognition memory is well-described (Miranda and Bekinschtein, 2018). But recognition memory necessitates a comparison of the encountered item with a previously formed record. Furthermore, item discrimination is not only determined by item form, and features, but also item size. Here, we explored to what extent the PRC is involved in de novo encoding of item dimensions and further investigated to what extent the different cortical layers of areas 35 and 36 might be functionally differentiated in this regard. To this end, we examined somatic immediate early gene (IEG) expression in the superficial and deep layers of areas 35 and 36 induced by novel exposure to small and discrete (microscale) and large and overt (macroscale) objects. IEG expression triggered by these events was detected using cellular compartment analysis of temporal activity by fluorescence in situ hybridization. Overall, our results indicate that the different cortical layers of area 35 engage in de novo encoding of item identity, in a dimension-specific manner. Whereas layers III and V-VI of the middle and caudal area 35 engage in de novo encoding of both kinds of item identity, superficial layer II, plays a subordinate role in the novel encoding of only macroscale item information. By contrast, in area 36, layer III is specifically involved in the de novo encoding of microscale items. Taken together these data suggest that areas 35 and 36 play distinct roles in the encoding of novel information about item dimensions.

These differences may derive from the anatomical structure of areas 35 and 36. From a histological point of view, there are very distinct differences in the cortical layers between areas 35 and 36. Area 35 is considered to be an agranular cortex (lacking layer IV) whereas area 36 has a weakly pronounced layer IV (Burwell, 2001). In addition, area 35 exhibits no clear separation of layers II and III in comparison to area 36 (Burwell, 2001). Moreover, areas 35 and 36 can be further divided into dorsal and ventral subcompartments and area 36 has a third segregation into a posterior subcompartment (Burwell, 2001). In the present study, we did not further differentiate these subcompartments. Nuclear Arc mRNA expression was scrutinized from layers II, III, and (I)V-VI of the middle and caudal PRC, and thus, included nuclei from the entire dorsoventral axis of either area 35 or area 36.

The anatomical connections along the rostrocaudal axis of the PRC and between areas 35 and 36 also exhibit several interesting differences (Furtak et al., 2007): compared to all other PRC levels, caudal area 36 receives the strongest input from visual association areas, and also the input from the POR, which receives even stronger visual inputs, predominantly targets area 36 rather than area 35 (Burwell and Amaral, 1998a; Furtak et al., 2007). Thus, area 36 may be better equipped to process information about microscale items than area 35. In addition to strong inputs from visual association areas, area 36 receives stronger polymodal associational inputs compared to area 35 (Burwell and Amaral, 1998a), suggesting that highly pre-processed sensory information reaches area 36. Thus, microscale item processing by area 36 may be supported not only by visual, but also tactile and odor inputs, for example.

Scrutiny of the intrinsic connections of the PRC also reveals very distinct features of area 35 compared to 36: Area 36 strongly projects to all rostrocaudal levels and the projections may target preferentially more ventral levels within this area (Burwell and Amaral, 1998b; Burwell, 2000). In turn, mainly ventral area 36 projects to the same rostrocaudal level of area 35, whereas area 35 makes only weak intrinsic connections and provides only weak feedback projections to area 36 (Burwell and Amaral, 1998b; Burwell, 2000). These differences in intrinsic and reciprocal connectivity of area 35 and 36 could suggest that area 36 may forward visual object information to area 35.

It has been reported that the PRC supports spatial memory by intrinsic information encoding or support of pattern completion (Ramos and Vaquero, 2005; Ramos, 2008, 2017; Barry et al., 2016). For example, PRC lesions result in impaired retrieval of spatial memories (Ramos and Vaquero, 2005; Ramos, 2008). In addition, several weeks after novel spatial learning, enhanced expression of Arc and Fos proteins is evident in the PRC (Barry et al., 2016). This involvement in long-term encoding contrasts with studies that report no engagement of the PRC in the learning of novel spatial arrangements of familiar stimuli (Wan et al., 1999; Jenkins et al., 2004; Aggleton et al., 2012), but may reflect a temporal aspect to the engagement of the PRC in these processes. Given the timeline of our study, we assume that the response of the PRC to novel macroscale or microscale object presentation relates more to the presentation of the item themselves than their spatial configurations (Aggleton et al., 2012). But the differentiated responses of area 35 and 36 to item dimensions introduces the interesting possibility that the PRC may be able to discriminate item size.

If the PRC does not support item-place information encoding per se, the question arises as to how the hippocampus acquires key information to execute this task itself. Anterograde and retrograde tracing studies have reported an indirect connection from layers II/III of the areas 35 and 36 to the hippocampus via the LEC layer II/III, but projections of area 35 are more intense than of area 36 (Burwell and Amaral, 1998b; Pinto et al., 2006; Doan et al., 2019). Moreover, electrical stimulation of layer II/III of area 35 results in postsynaptic potential in the neighboring LEC, supporting monosynaptic inputs to layer II/III of LEC originating mostly from layer II/III of area 35 (Doan et al., 2019). These are likely to play a key role in the processing of novel object memory. Consistent with these anatomical observations, layer III of both middle and caudal area 35 exhibits an increase in activity as a consequence of both microscale and macroscale item encoding. This result indicates that neurons in this layer of area 35 process novel object information independently of the dimensions of the objects.

The hippocampal CA1 region plays an important role in the learning about microscale item-place configurations (Kemp and Manahan-Vaughan, 2004; Hagena and Manahan-Vaughan, 2011; Hoang et al., 2018), and the microscale and macroscale behavioral paradigms used in the current study trigger both synaptic plasticity and somatic IEG expression in distinct subcompartments of the hippocampus (Hoang et al., 2018; Hoang and Manahan-Vaughan, 2021). Interestingly, it has been reported that a direct projection from neurons in the superficial layers of the PRC to the CA1 region and more strongly to the subiculum of the hippocampus exists (Naber et al., 1999; Furtak et al., 2007; Agster and Burwell, 2013; Suter et al., 2013). In our study, we observed that layer III of area 36 was activated only by the exposure to novel microscale items, and unaffected by macroscale items. This finding prompts the possibility that the engagement of layer III of area 36 in microscale object encoding and the weak direct connection from the PRC to the hippocampus (Furtak et al., 2007; Agster and Burwell, 2013) may support microscale item-place encoding in the hippocampal CA1 region.

The hippocampal dentate gyrus region engages in macroscale item-place encoding (Kemp and Manahan-Vaughan, 2008; Hoang et al., 2018, 2021). This prompts the question as to whether this process may be supported by macroscale item encoding in area 35. In contrast to the direct connections of PRC with the CA1 region, tracing studies have generated conflicting results with regard to a possible direct connection from the PRC to the dentate gyrus: Several anatomical studies indicate that the dentate gyrus does not receive projections from the PRC (Kosel et al., 1983; Mcintyre et al., 1996; Naber et al., 1999), whereas others argue that the PRC/entorhinal cortex inputs terminate on in the middle third of the molecular layer of the dentate gyrus, which can be correlated to the medial perforant path input (Canning and Leung, 1997; Canning et al., 2000). However, a more recent study suggests that a monosynaptic input extends from the superficial layers of the PRC at least to newborn granule cells (Vivar et al., 2012). Lesioning this particular pathway results in impaired pattern separation (Vivar et al., 2012), indicating that an interconnection between the DG and the PRC may exist. In the current study, neurons in layer II of the PRC exhibited only sparse activity in response to novel object learning. These neurons, in particular in the caudal compartments of area 35, only become active when animals engage in the exploration of novel macroscale, but not microscale items. These results suggest in particular that activity in layer III of area 36 and to a smaller extent also layer II of area 35 may be input-specific. Keeping in mind that the PRC mainly indirectly projects to the hippocampal formation, it is likely that activation of the superficial layers II and III may be correlated with an encoding of novel object information that is transmitted via the LEC to the hippocampus. Here, information about novel objects might be projected from PRC superficial layers to the LEC that in turn sends information either directly or indirectly to the related hippocampal compartments for the encoding of the spatial components of the object configuration.

In our study, neurons in the layers V-VI of area 35 encoded both kinds of items presented, suggesting that item dimension is not discriminated in these layers. It has been reported that the back projections from the CA1 region/subiculum preferentially target the layers V-VI of the caudal PRC (Swanson and Cowan, 1977; Deacon et al., 1983; van Groen and Wyss, 1990; Kloosterman et al., 2003). Thus, another interesting possibility is that the hippocampus may send general information about object dimensions and/or item-place both directly and indirectly via the LEC back to the PRC (Burwell and Amaral, 1998b). This information may then be encoded in layer V-VI of area 35 (see: Figure 6). In order to find out if activity in these layers may reflect novel encoding or recognition/retrieval (Burke et al., 2012a), and/or spatial components of object configurations (Ramos and Vaquero, 2005; Ramos, 2008), further investigations will be necessary.

Our study indicates that areas 35 and 36 are functionally differentiated with regard to their role in novel item dimension encoding. Differences in the anatomical connections of areas 35 and 36, and differences along their rostrocaudal levels indicate that these structures are also functionally differentiated with regard to object recognition memory (Burwell and Amaral, 1998a, b; Furtak et al., 2007; Agster and Burwell, 2013). Empirical evidence for this, in particular in rodents, is sparse but findings indicate a differentiated involvement of areas 35 and 36 in both object recognition memory and novel item encoding (Fujimichi et al., 2010; Sethumadhavan et al., 2020). Most of the studies that examined the PRC used substance injections or lesions of the PRC, to understand the role of the PRC in object recognition memory (Mumby and Pinel, 1994; Winters and Bussey, 2005; Albasser et al., 2009; Aggleton et al., 2010). These methods, although useful in gaining an understanding of the role of PRC in object recognition and related behavioral tasks, offer limited potential for the functional differentiation of area 35 and area 36.

The IEGs, Arc and cfos, can be used as biomarkers of neuronal activity and synaptic plasticity (Zhu et al., 1995b; Guzowski et al., 1999, 2000; Burke et al., 2012a). Most studies examined Fos expression in the PRC to determine the regional distribution of activated neuronal populations in conjunction with object recognition (Zhu et al., 1995b; Albasser et al., 2010, 2013; Seoane et al., 2012). Novel object learning in darkness increases Fos protein levels in the rostral areas 35 and 36, indicating its role in darkness perception (Albasser et al., 2013). In contrast, novel object learning in illuminated conditions enhances Fos expression in the caudal areas 35 and 36 (Zhu et al., 1995b; Albasser et al., 2009, 2010, 2013). Interestingly areas 35 and 36 of the middle compartment of the PRC exhibit similar levels of Fos protein for the novel object exposure and object recognition groups (Albasser et al., 2010, 2013). However, none of these studies distinguished Fos expression between the layers of the PRC (Zhu et al., 1995b; Albasser et al., 2010, 2013; Seoane et al., 2012). The expression of Fos protein peaks approximately 90–120 min after the start of an experience (Kovacs, 1998; Zangenehpour and Chaudhuri, 2002). This imprecise onset of transcriptional activation of the Fos protein results in a low temporal resolution of Fos protein expression, which can also confound interpretations of IEG expression related to a learning event.

In contrast to the Fos protein, the activation of Arc gene expression by a learning event is rapid and dynamic with the peak expression of Arc mRNA in cell nuclei occurring within 5–6 min of commencing an exploration period (Guzowski et al., 1999; Guzowski, 2002). This approach has the advantage of excluding that events that occurred before or after the exploration period do not inadvertently affect gene expression. This, in turn, allows a more precise mapping of activated neurons using Arc mRNA compared to Fos protein. Only a few studies have used somatic Arc expression to scrutinize the role of the PRC in object memory. Nonetheless, blocking Arc expression in the PRC of rodents, leads to an impairment in the differentiation of similar object representations, without affecting spatial, or distinct object representations (Miranda et al., 2017). These findings align with studies supporting the role of the PRC in resolving feature ambiguity (Bartko et al., 2007). Another study of Arc mRNA expression in layer V of areas 35 and 36 (Burke et al., 2012a), compared the effects of a novel context using the same object constellation twice (Burke et al., 2012a). The results indicated that novel and familiar exploration of objects results in enhanced Arc mRNA expression in layer V (Burke et al., 2012a), but interestingly a difference in Arc expression between areas 35 and 36 was not detected. In the present study, we detected a differentiated engagement of areas 35 and 36 in novel object learning, suggesting an area and layer specific activity of the PRC in novel object encoding that is related to item dimensions.

Comparing our behavioral task with the behavioral approach used by others reveals that most groups presented large-scale objects that were placed on the floor of an arena (Winters et al., 2004; Balderas et al., 2008; Albasser et al., 2009, 2010; Aggleton et al., 2010), that in many respects reflect our use of macroscale items. In some other studies, objects only could be explored visually through holes because they were presented indirectly behind one-way mirrors in paired-viewing tests (Zhu et al., 1995a,b, 1996), similar to our microscale approach where small items were placed in holes on the floor and could only be seen if the animals poked their noses into the holes. Some of these studies used objects with different dimensions (largest dimension: 3–15 cm; Zhu et al., 1995b) or height (5–20 cm; Winters et al., 2004), but the behavioral and neuronal response to different scales of objects was not explicitly examined (Zhu et al., 1995a,b; Winters et al., 2004). Interestingly in “paired-viewing” studies, it was reported that the use of “2D” pictures that were passively viewed on a screen instead of “3D” objects that were actively explored, generates a similar PRC encoding response: enhanced Fos levels can be detected in the PRC only for single novel pictures (not constellations) or novel objects compared to the re-exposure to either condition (Zhu et al., 1995b, 1996; Wan et al., 1999; Seoane et al., 2012).

It is important to emphasize that our study explored IEG expression triggered by novel object exposure. Most other studies compared IEG expression between a novel exposure and re-exposure (item recognition) state (Albasser et al., 2010, 2013) in the absence of a naïve control. The difference in our findings compared to the findings of these studies may thus serve to highlight how different compartments of the PRC support novel item processing, as opposed to item recognition. Along these lines, it is possible that there is no clear border for an involvement of the middle, and caudal compartment in novel visual object recognition memory, rather a graded change may occur from anterior to posterior in the strength of involvement of the PRC in this process. By contrast, a differentiated involvement of areas 35 and 36 in novel item memory is evident that is based on item dimensions.

The duration of exposure to the items may also play a role, however. We presented a constellation of three different objects for a duration of 5 min. Burke and colleagues (Burke et al., 2012a) also used a 5-min exploration period, but in contrast to our study, they presented five different objects. Due to a decrease in exploration time in the second exposure, they concluded that the animals remembered the objects presented (Burke et al., 2012a). A study using the same behavioral tasks that we used revealed a longer exploration time of the novel compared to the familiar object, confirming that the animals remembered the previously presented objects (Hoang et al., 2018). The difference in the number of objects is probably not the reason why we detected no change in IEG expression in area 36 in layer IV-VI in response to novel object recognition, in contrast to the similar response of layer V of areas 35 and 36 reported by Burke et al. (2012a). It is more likely that the results differ because we not only analyzed layer V but also included nuclei of layers (I)V-VI in our analysis.

Taken together, the results from our current study suggest a layer- and area-specific involvement of the PRC in the processing and encoding of information about novel large- and small-scale objects. Whereas area 35 exhibits an increase in neuronal activity induced by object learning that is generally independent of object dimension, area 36 only engages when animals explore less prominent microscale items. Furthermore, differences in object dimension trigger layer-specific gene encoding in the PRC: neurons in layer III of area 36 are only involved in learning about microscale items, suggesting that distinct afferent and efferent connections may be responsible for the integration of less prominent item information. By contrast, the superficial layer III together with layers V-VI of area 35 are activated by both types of object dimensions. In conclusion, these findings suggest that areas 35 and 36 are functionally specialized to enable item encoding on the basis of item dimensions.

The raw data supporting the conclusions of this article will be made available by the authors, upon reasonable request.

The animal study was reviewed and approved by Landesamt für Arbeitsschutz, Naturschutz, Umweltschutz und Verbraucherschutz, Nordrhein Westfalen.

DM-V and CS devised the concept and experimental strategy of the study. Experiments were conducted by NS and CS. Data analysis was conducted by NS, CS, and T-HH. DM-V wrote the article, with contributions from all authors. All authors approved the submitted version.

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.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.