In the present study we investigated how the perceived temporal distance (PTD) of events can affect the Mental Time Travel (MTT) ability and the subtending neural activations. We systematically compared the PTD effect for relative-past and relative-future events in a continuous manner, based on participants’ estimated PTDs, and we analysed the BOLD signal changes as a function of these estimates. Our results showed that reaction time (RTs) and accuracy are linearly related to PTDs for relative-past events in the Present and for relative-future events in the Past. This finding suggests that, even without explicit instructions, participants implicitly represent the temporal distance of events while performing the task. It is noteworthy that the relative-past events in the Present and the relative-future events in the Past can represent the same “segment” of the putative Mental Time Line (MTL). Within this segment, the spatial representation of Past influences event processing, including the PTD effect on RTs, based on the self-location adopted in time (resulting in a similar effect for relative-past events in the Present and relative-future events in the Past). Moreover, PTDs correlate with the accuracy for relative-past events in the Present and the accuracy for the relative-past Self-Reference, regardless of the Self-Projection condition. These results suggest that the PTD effect we observed is stronger for the past: the closer the past events, the slower and less accurate the performance. The regression analyses revealed a distance effect for PTDs in relative-past and relative-future events in the Past, as well as in relative-past events in the Present. This effect persisted for both relative-past and relative-future events when collapsing the Self-Projection conditions. Overall, these findings indicate that a decrease in PTDs results in an increase in RTs, i.e., the closer the events, the slower the reaction time, suggesting a causal linear effect of PTDs on RTs. In previous literature, the distance effect is usually taken as evidence for a continuous representation of magnitudes (Pinhas et al. 2012), which in our case refers to the representation of PTDs on the MTL. Considering previous studies which focused on a-priori dichotomous categorization to study the effect of temporal distance in MTT, our contribution to the MTT literature is especially valuable. Indeed, in the current experiment, PTDs provided by participants did not reproduce a dichotomic evaluation of distances but showed a continuous distribution. Notably, we suggest that these PTDs, rather than the a-priori dichotomic categorization, represent a better predictor of the MTT performance.

It could be pointed out that we did not find the same temporal distance effect for the Future as in previous works (Arzy, Adi-Japha, et al., 2009a; Gauthier and van Wassenhove 2016a, b). This difference could be ascribed to methodological differences between previous studies and the current one. In those earlier works the authors provided participants with the exact date (in terms of years) of the relative-future events that participants had to retrieve when executing the task. In our task, instead, participants had to implicitly estimate the date of each future event which was not a-priori established. If the exact date for relative-past events is already stored in the long term memory and the order of such events are well established on MTL, this is not the case for relative-future events. This difference is crucial and could have masked the perceived temporal distance effect related to the Future, suggesting dedicated mechanisms involved in future MTT. Indeed, to estimate PTDs of future events, cognitive functions, such as anticipation, simulation, mental imagery as well as scene construction processes, are involved. Hence, the heterogeneity of these processes may have prevented the emergence of a future-related temporal distance effect. Anyhow, future research is needed to verify this intriguing hypothesis.

From a neurocognitive perspective, perceiving events as closer results in longer RTs during the MTT task, owing to the activation of more complex representations, i.e., representations rich in episodic and semantic details, which require longer time to be recalled or imagined. Such complexity level decreases as the chronological distance from the present increases (D’Argembeau and Van der Linden 2004). These detailed cognitive representations are tied to considerable brain activity, predominantly in the medial temporal lobe. The hippocampus retrieves and integrates an array of memory details, including sensations, emotions, spatial and contextual details (D’Argembeau, 2020), as well as their temporal order (Gauthier et al. 2020). In particular, the anterior and posterior regions of the hippocampus display different involvement depending upon the detail level, following a gradient from coarse (anterior) to fine details (posterior). In addition, the hippocampus combines segmented information from primary sensory regions and builds event models in the posteromedial cortex (precuneus and posterior cingulate/ retrosplenial cortices) and angular gyrus (D’Argembeau, 2020). Finally, it works alongside parietal areas, particularly inferior parietal cortices, contributing to the vividness of memory recall and allocating attention during this process (Ciaramelli et al., 2010; Cabeza et al. 2008). The complexity of representations is also related to their spatial, temporal or interpersonal distance. According to the Construal Level Theory by Trope and Liberman (2003), closer objects, including events, are represented more concretely and through more detailed “low-level construals”, including specific knowledge, as well as contextual features. This contrasts with more distant events, which are represented more abstractly. From this perspective, the concrete representations tied to proximate events are associated with heightened activity in the anterior and dorsal regions of the medial prefrontal cortex.

The PTD effect here reported is similar to the distance effect observed in the numerical domain (Moyer and Landauer 1967), besides in other temporal tasks (Bonato et al. 2016). The existence of a common mechanism for quantities and distances estimation, be they spatial, temporal, or numerical in nature, has been already suggested, and it implies that the closer an item is to a certain point of reference, the longer it takes to determine the correct answer (for a review see Bonato et al. 2012). In the temporal domain, Bonato and colleagues (2016) observed that right brain damaged patients with left neglect were also impaired in ordering events in time: they exhibited slower responses to items that occurred before the temporal reference provided in the experimental manipulation. Their finding adds further evidence of a common origin of the spatial effects characterising both the numerical and temporal representations of order. In the numerical domain, this distance effect is typically observed in magnitude judgement tasks, where responses become progressively slower and less accurate as the numerical difference between two numbers decreases. van Dijck and Doricchi (2019) demonstrated an asymmetrical numerical distance effect in right brain damaged patients with left spatial neglect. These patients were abnormally slow only when responding to the closest smaller number as compared to the referential number (i.e., 4 when reference is 5). However, patients with spatial neglect showed a normal SNARC effect (Spatial-Numerical Association of Response Codes, i.e., automatic response association between small/large numbers presented on left/right space, respectively - Dehaene et al. 1993), and performed well in parity judgements (“2 is an odd or an even number?”). In the light of this, van Dijck and Doricchi (2019) proposed that spatial neglect does not affect the spatial coded response selection. Namely, the parity judgments can be solved by activating over-learned representation of numbers on the Mental Number Line. Building on this hypothesis, the perceived temporal distance effect reported in this study for the past could rely on an “over-learned” or “well-stored” representation of temporal orders on the MTL, allowing the PTD effect to emerge. Conversely, the lack of defined knowledge about the future could lead to an “interference” of the envision processes of possible scenarios, affecting (or abolishing) the spatialization of the temporal order of relative-future events and, consequently, the PTD effect in the Future.

To the best of our knowledge, our study represents the first attempt to systematically evaluate the modelling effect of PTDs on the BOLD signal in a continuous manner. We found that the perceived proximity of relative-past and relative-future events is associated with the activation of a shared network, encompassing bilateral angular gyrus, temporal and parietal areas, which include the temporo-parietal junction (TPJ, Geng and Vossel 2013), retrosplenial cortex, middle and superior frontal gyri. Our subsequent parametric analysis incorporating RTs as factors confirmed that the activations observed as a function of PTDs are indeed specific to the PTD effect and not influenced by an unspecified time-on-task effect, as other distinct regions are linearly related to the RTs.

Regardless of the specific task employed, a “core network” mediating both the retrieval of relative-past and the envisioning of relative-future events has been described, comprising brain areas very similar to those reported in our results (Arzy et al. 2009b; Benoit and Schacter 2015; Hassabis et al. 2007). This aligns with the recent proposal by Addis (2020) of a single “simulation system”. Notably, our findings revealed the involvement of these areas regardless of whether the events were judged as “past” or “future”. This supports our prediction of a “PTD core network” activated for the processing of perceived temporal distances, especially involving TPJ, inferior parietal lobule, angular gyrus, frontal and posterior areas. This is consistent with Arzy et al. (2009), who identified TPJ as a key structure for the encoding of the self in both the temporal and in the spatial domain. It also aligns with Parkinson and colleagues (2014), who demonstrated that the representations of egocentric spatial, temporal and social distances converge in right TPJ, which is involved in the self-other distinction and in the mental representation of space and events along the Mental Time Line. In addition Gauthier and van Wassenhove (2016b) found a common representation of distances in right IPL/AG and anterior insula both for the temporal and the spatial domains. Right IPL is identified as a pivotal area for egocentric re-mapping and computation of distances in both domains, as well as for the perception of temporal order. Consistently with these findings, Peer et al. (2015) tested distances (close vs. far) in time, space and personal relationship domains, and demonstrated a common activation in the precuneus, IPL, and medial prefrontal cortex. This activation was explained by the processing of the distance between the self-location and the cued stimulus. This observation aligns with the numerical literature, which highlights the crucial role of the posterior parietal cortex, specifically the intraparietal sulcus and angular gyrus, in numerical representations and manipulations. Notably, the left angular gyrus’ involvement in number processing may be linked to the linguistic foundation of arithmetic computations: as this region is also involved in various visuospatial tasks, including eye and/or attention orienting, mental rotation, and spatial working memory, it was identified as a “common ground” for both numerical and spatial domains (Dehaene et al. 2003). Thus, the posterior parietal cortex could also bind numerical and temporal domains.

Our results showed that the brain areas mediating the perceived temporal distance for relative-past and relative-future events do not overlap entirely. This suggests that some of the involved processes could differ, in agreement both with previous literature and with our predictions. It is worth noting that the brain network associated with PTD partly overlaps the Default Mode Network (DMN), which is known to be involved in the self-referential and internal processing (Buckner and Carroll 2007) and in mental travel, allowing for the change of the mental location of the self and the reorganisation of one’s surroundings (Hayman and Arzy 2021). Notably, two subsystems were identified within the core network of the DMN (Addis et al. 2009). The first is the remembering subsystem, activated only during the retrieval of detailed past events (i.e., hippocampus, parahippocampal gyrus and widespread regions of posterior visual cortex; Gaesser & Addis, 2011; Thakral et al. 2017). The second is the imagining subsystem, active when envisioning future scenarios (i.e., anterior hippocampus and widespread medial prefrontal and parietal regions; Addis et al. 2007, 2009). Furthermore, neuropsychological studies revealed that right-brain damaged patients exhibiting left neglect were slower when responding to relative-future than to relative-past events, owing to their spatial working memory deficit (Anelli et al. 2018). Additionally, patients with lesions in the ventro-medial prefrontal cortex were specifically impaired when projecting themselves towards the future and when judging the future location of events on the MTL (Ciaramelli et al. 2021). Furthermore, it has been reported the case of a patient whose gray matter volume reduction in thalamus, fusiform gyri and cerebellum bilaterally, was related to his retrograde amnesia, but also with his impairment in envisioning future events (De Luca et al. 2018). Interestingly, we found that the perceived proximity of relative-future events engages left parahippocampal and lingual gyri and right cerebellum. This is in line with our hypothesis of an additional involvement of processes, such as attention, visual imagery and cognitive resources, in constructing possible future scenarios, when ordering events on the MTL.

The parahippocampal cortex (PHC) is implicated in autobiographical memory retrieval, prospection, navigation, Theory of Mind and Mental Time Travel (Hayman and Arzy 2021; Spreng et al. 2009). Recent studies suggest that PHC also plays a role in constructing scenarios that are alternative to reality (DiNicola et al. 2023). In this regard, Irish and colleagues (2015) found that the gray matter volume of PHC correlates with the ability to construct spatially coherent scenes, contributing to the processing of spatial and contextual associations. Moreover, Epstein and colleagues (2003) demonstrated that the parahippocampal place area (PPA) enables the computation of the location and the orientation of the self with respect to the internal map. In our results, PHC could mediate the representation of future scenes based on the temporal distance of events, envisioning them in spatial terms relative to the location of the self.

Together with the hippocampal and parahippocampal gyri, the lingual gyrus has been identified as a key region mediating creativity and divergent thinking (Dietrich 2004; Gilbert 2001). Namely, Jung et al. (2010) observed that the thinner the gray matter volume of the lingual gyrus, the lower the scores in divergent thinking tasks, particularly in the ideational fluency (i.e., the quantity of original ideas provided). Expanding on these findings, Zhang et al. (2016) demonstrated that a larger volume of the lingual gyrus is associated with increased creativity, as well as with enhanced cognitive flexibility. Furthermore, Zhang and colleagues (2014) highlighted the involvement of the left lingual gyrus in processing relevant visual imagery during the generation of inventive ideas. Additionally, research by Slotnick and Schacter (2006) revealed that the left lingual gyrus plays a role in spatially specific memory processes, implicitly encoding spatial information related to stimuli positioned on the right portion of the screen. Our results could combine all the presented roles of this brain area in the generative process of envisioning future scenarios, since divergent thinking involves retrieving knowledge from memory to creatively organise mental representation as a new idea (Zhang et al. 2020). We speculate that the left lingual gyrus may engage in spatial memory processes to create original future scenarios (located on the right portion of the Mental Time Line), particularly when an implicit representation of the temporal distance of perceived events is considered, as in the MTT task.

Functional MRI studies revealed increased activations in the cerebellum during the processing and the construction of future events (Addis et al. 2007), during episodic future thinking (Szpunar et al. 2006), and when predicting future action sequences during mentalizing tasks (Van Overwalle et al. 2022). Interestingly, Oliveri and colleagues (2009) found that repetitive Transcranial Magnetic Stimulation (TMS) on the right cerebellum of healthy participants affected their speed in responding to the future tense of action verbs. In addition, these authors proposed a right cerebellar-left motor brain network involved in anticipating future events (Oliveri et al. 2009). This hypothesis was confirmed by neuropsychological evidence from patients with cerebellar lesions, whose ability to predict, anticipate and reconstruct sequences of events was impaired (Leggio and Molinari 2015). Furthermore, the functional connectivity between the cerebellum and the mentalizing network (i.e., angular gyrus, parahippocampal gyrus, lateral occipital cortex, middle temporal gyrus and precuneus) was altered in patients with the behavioural variant of Fronto-Temporal Dementia with deficits in social behaviour (Olivito et al. 2022). This evidence strongly supports the role of the cerebellum in generating alternative scenarios based on internal models or past experiences (Oliveri et al. 2009; Schacter et al. 2007). Additionally, our findings suggest the involvement of the right cerebellum in constructing the future temporal location and anticipating relative-future events in the context of the MTT task, especially in relation to the perception of temporal distances. It is plausible that activation of the right cerebellum may facilitate collaboration among brain areas within the PTD core network. This collaboration could be crucial for generating original future scenarios, involving the lingual gyrus for visual processing and PHC for adopting a different spatial perspective. Nonetheless, future research is needed to test the model we propose here, better exploring the co-activation pattern and the connectivity between these regions.

Our findings assume added significance when considered in the context of the hypothesis by Gauthier and van Wassenhove’s (2016b), who proposed that the brain distinctly represents temporal and spatial egocentric distances, as evidenced by the activation of an extensive network specific for the spatial representation of proximity. This network comprised the precuneus/retrosplenial cortex and superior parietal lobule, inferior parietal sulcus, right superior frontal cortex, pre–supplementary motor area, rostrolateral prefrontal cortex, inferior temporal and parahippocampal cortices and left cerebellum. Our study revealed a strikingly similar neural network, however here this network is involved in processing perceived temporal distances. Using both temporal and spatial tasks, Gauthier and van Wassenhove (2016a, b) instructed participants to memorise various details about the events, therefore their temporal task did not involve a spatial component, which was instead prevalent in our study. It is to be noted that, for the first time, in the current study the projection of the self and the localisation of the events on the MTL were executed “on-line” when participants were tested, whereas temporal locations were provided before executing the tasks in Gauthier and van Wassenhove studies (2016a, 2016b). Consequently, the observed similarities between our findings and the activations identified in Gauthier and van Wassenhove study (2016b) for spatial distance computation may be attributed to the spatialization of time induced by our task, which emphasised the spatial representation of time. To support this hypothesis, we selectively masked our neural activity associated with the general MTT process and found that the temporo-parietal, retrosplenial, temporal medial and frontal areas resisted, suggesting that they are specifically involved in the perceived temporal distance computation.

In conclusion, our study provides insights into the mechanisms underlying the processing of subjectively perceived temporal distances in the MTT task, thereby enhancing the ecological validity of the task. The functional imaging findings show posterior parietal, temporal and frontal areas subtending the PTD effect, a network which appears very similar to the one engaged in spatial distances processing. The behavioural results also reproduce spatio-temporal effects related to the distance of events. Thus, our study supports the hypothesis of a common cognitive representation between space and time, as suggested by the ATOM theory (Walsh 2003b; Bueti and Walsh 2009). According to an alternative yet interesting hypothesis, time and space (along with numbers) might interact with each other only at the functional level of working memory. This is suggested by the observation that all three dimensions - time, space, and numbers - require spatial attention and serial ordering within working memory to accomplish tasks (van Dijck et al. 2013). In the same vein, during the MTT task, working memory may build a spatial representation of time, by guiding attention towards long-term memory representations of space and time, thus facilitating the retrieval of past and the imagination of future events. Future research will be needed to disentangle the hypothesis of the spatial representation of time from the spatialization of temporal processes in the working memory.

Our study provides an initial exploration of the role of perceived temporal distance in Mental Time Travel, although our experimental design prevents us from directly comparing the effects of objective and perceived temporal distances on MTT and their associated neural correlates. Future studies, specifically designed to directly contrast these two types of temporal distances, are necessary to enhance our understanding of their implications in MTT and of time perception.