Introduction

The ability to keep track of events unfolding over time is a fundamental and instinctive feature of humans and animals’ perception and cognition, essential for adaptation to complex and challenging environments1. Indeed, predicting the events to come, interacting with others, crossing the road, remembering a planned appointment, or catching the train on time are all activity of everyday life supported by temporal processing, without which, autonomous behaviors would not be possible. Temporal processing encloses a wide range of timescales (i.e., from milliseconds to decades), across multiple sensory modalities. Nevertheless, the temporal experience could be considered as formed from four main dimensions: temporal order and simultaneity, the processing of durations, the feeling of the passage of time and mental time travel2. These dimensions are usually investigated in laboratory using various temporal tasks and questionnaires (detailed in Box 1). They call on partially shared brain underpinnings and have close links with most cognitive processes3,4. Nevertheless, the debate upon which neural mechanisms underlies the processing of time is still open (Box 2).

Evidence agrees that temporal sensitivity5,6,7 and semantic temporal concepts develop until adulthood8,9 and then become more variable and less precise with advancing age10. Yet, there is no consensus about which biological and psychological processes evolve and support this development. While some data suggest changes of the speed of the internal clock11,12,13,14, others show that the evolution of timing abilities, duration processing in particular, would be supported by the development of working memory15,16,17,18,19, executive functioning20,21,22,23 and attention24,25,26. However, the involvement of cognitive processes depends on the temporal tasks used and whether they are more or less cognitively demanding27,28. In addition, although several compelling works reviewing evidence of temporal abilities in new-borns29, infants30, children6,7,9,31, and older adults10,32,33 exist, a unified trajectory addressing the development of the underpinnings of the ability to perceive and organize the time from the intra-uterine life to older age is lacking.

The present neurodevelopmental perspective claims that our ability to perceive and prospectively process durations and time passing emerge thanks to the interaction between various endogenous (e.g., brain maturation) and exogenous factors that we are exposed to, depending on the different stages of life. According to this perspective, the temporal processing, among other factors7, evolves and refines itself step by step at each life stage, as a function of the gradual shaping of the structural and functional architecture of the brain (Fig. 1). This approach could help to specify the maturation of the regions associated to temporal processing (Box 2) throughout lifetime while highlighting its evolutionary contribution. What is more, retracing the stages of a physiological course of timing abilities will add important bricks in the building of a neurodevelopmental model of human time processing, which will be proposed in the end part of the review (Fig. 2).

Fig. 1: Lifetime lines.
figure 1

Life-span trajectories of structural and functional development of the human brain paralleling the emergence of each of the four aspects of temporal processing. Created with Biorender.com.

Full size image
Fig. 2: The neuroanatomical development of the temporal processing circuit.
figure 2

The proposed additive maturation model of RDPM, according to the functional specialization of the temporal processing circuit, for each of his aspect. Red: Rhythms; Yellow: Durations; Green: Passage of Time; Purple: Mental Time Travel. Created with Biorender.com. Brain templates have been reconstructed with Brainstorm software by Tadel et al. (2011)294 leveraging on Li et al. (2015)295; Shi et al. (2011)296; Fonov et al. (2009)297; Dong et al. (2020)298.

Full size image

Far to conduct an exhaustive review of the entire cerebral and cognitive development, we will provide an overview of the emergence of the ability to perceive and process time, in parallel to the development of the brain structural and functional milestones considered to support prospective judgments in explicit temporal processing tasks. Furthermore, we will also focus on the rise of the cognitive functions most related to these temporal skills. For each life stage, we will highlight the adaptive advantages provided by the temporal dimension which unfolds. Lastly, we will show that deviations from the healthy neurodevelopmental trajectory could result in an impaired ability to keep trace of time (Boxes 3 and 4).

The intrauterine life (15 Post Conceptional Weeks (PCW) – 0 years old)

Starting from the second trimester, the formation of neural tube (neurulation) as well as the generation of neurons (neurogenesis) are for the most part completed. Since the 15th Post Conceptional Week (PCW), the brain structure starts to assume its human-specific configuration, with the ganglio-thalamic formation connecting the telencephalic proliferative zone to the major thalamic regions34,35,36. The formation of the thalamus-based relay of information from the 24th PCW has a pivotal role in the setting up of cortical networks, as it allows to sensory inputs to reach the developing cortex37, laying the groundwork for the formation of sensorimotor, visual, and auditory circuits. Since the womb life, fetuses are exposed to several internal biological rhythms, originated by the mother’s cardio-vascular system, breathing, walking, but also some stimuli of the external environment as sounds or the mother’s voice.

The setting up of the cortico-striatal pathways at this stage, as a result of the increased fractional anisotropy (FA) of the basal ganglia38,39, would set the ground for the emergence of an elementary form of sensorimotor synchronization. Near term fetuses indeed are able to perceive rhythmic sequences like maternal rhythmic movements, walking, rocking, or breathing and to coordinate with them accordingly. Evidence from studies using Heart Rate (HR) demonstrated that fetuses changed their own HR depending on the mother’s rhythmical movements40, and differentiated their breathing pace in response to rhythmical rocking or sinusoidal oscillations29. Also, fetuses react to auditory stimuli, such as variation in music and speech, with a prevalence of left temporal activation41. Studies have shown that fetuses preferentially react to the sound of mother’s heartbeat42 and voice43,44. Therefore, fetuses are already able to memorize and recognize the temporal harmonic structure of the prosody of the mother’s speech45.

Fetuses also produce rhythmic activities (kicking, sucking). Although they are fairly stereotyped and inflexible, some changes in rhythm production in response to the exogenous factors during fetal life have been noted by studies on finger sucking, which are observable by the 15th PCW and rhythmic mouth movements, from the 20th PCW29. The ability to adapt their own rhythm to an external one is crucial for the newborn to comprehend the temporal organization of interpersonal interactions, setting the stage for speech and communication46,47. Furthermore, biobehavioral synchrony plays a crucial role in mother-infant attachments, providing a constant safety link, and contributing to the development of the sense of self and relatedness48. The prenatal temporal competencies are therefore orchestrated by spatiotemporal synchronization, and appear in the abilities of fetuses to perceive, predict, and adjust behavior to external rhythmic structures.

Infancy (0-3 years old) – kindergarten years

The exponential formation of new synapses occurring at birth is followed by an extended period of synaptic pruning49. The myelination in the basal ganglia (caudate, putamen, globus pallidus) is paralleled with a rapid gray matter thickening50,51 in primary sensory and motor cortices. On the other hand, the proliferation and widespread migration of the new, mostly inhibitory, interneurons, to the cingulate cortex and superior frontal gyrus (SFG) are responsible of the gray matter growth in frontal areas52 occurring at around two years old53. Resting state functional connectivity studies showed that the primary sensorimotor and auditory networks are already formed at the time of birth54,55. Conversely, the connectivity of the visual network steeply increases during the next three months while fronto-parietal attentional (FPN) and default mode network (DMN) regions mature later56.

Neurophysiological footprints of speech perception are detectable as early as the 4th day of life, as shown by a greater activity of left auditory association cortices to dichotically presented syllables57,58. This increased left-ward activation has been demonstrated to be responsible of fast rhythms processing59, characterizing the phonemic rate of the speech60,61,62. Duration processing abilities have also been observed in the first months of life. As early as 4 months old, Provasi and colleagues (2011)63 observed that babies exhibited an adult-like psychophysical response to an adapted temporal bisection task, during which they were trained to look to the left after the long standard (1500 ms) and to right after the short standard duration (500 ms). The proportion of long responses, measured as the duration of the gaze, increased with the target duration, already demonstrating a relative sensitivity to time.

Successful duration discrimination to an auditory oddball-timing of milliseconds to seconds range intervals that differed by 1:2 ratio between the inter-stimulus interval standard (1500 ms) and the deviant stimuli (i.e., 370, 500, 750, 1000 ms) was observed in six months old babies64. Precision in interval discrimination to milliseconds to seconds range has also been demonstrated to increase from six to ten months, as shown by the larger amplitude of the Mismatch Negativity (MMN) and the increased ratio between the standard and the deviant interval65. Also, precursors of the P1–N1–P2–N2 complex sensitive to differences in tone durations have been observed in the left temporal and the premotor cortices in six to eight months old infants66.

Later, cortical thickness increases rapidly in superior temporal gyrus, parietal, and pre-motor cortices, and in the insula67. This cortical expansion, the refinement of primary sensory networks, together with the increased FA of commissural fibers could set the stage for temporal production, temporal comparison, and temporal discrimination abilities on short duration range (i.e., up to two seconds) across multiple perceptual modalities68,69.

The intrinsic connectivity reported by fMRI studies between regions of the dorsal attention, the salience, executive control network, in nine to twelve months old babies70, likely supports the development of cognitive abilities such as attention71 and visual-spatial skills72. This may explain the higher sensitivity to duration in the auditive modality than in the visual modality19,73,74.

For methodological reasons there are few studies on children aged between one and three years old concerning both the perception and production of duration and rhythm. The most basic motor actions following a rhythmical pattern such as pace tapping, stepping, and dancing are also followed by more complex activities like singing or playing music. As it will be mentioned in the following section, there is still an improvement of these abilities between three and six years based on the development of motor skills and motor inhibition. The ability to detect temporal regularities of prosodic cues help to grasp acoustic statistical dependencies between words75,76 fundamental for language acquisition77.

Late Infancy (3–6 years old) – Pre-school years

During early childhood, the myelination of the cortico-cortical association tracts is coupled with cortical thickness decreases within medial and polar occipital regions, and to a greater extent in prefrontal and parietal cortex78. In parallel, VBM studies conducted in toddlers showed gray matter volume increases of some prefrontal and cerebellar regions, which decreases consistently in volume thereafter79,80. From a functional perspective, hubs of the DMN are already present but in an immature form81, failing to synchronize into a coherent network82. Yet, a significant coupling of dorsal mediofrontal cortex and posterior regions was observed as early as age four at rest83. Also, throughout early childhood, the bilateral precentral gyrus is pivotal for functional network development, showing the major nodal efficiency increase in the development of structural networks84. The increasing network organization emerging by this stage, could explain the differences in temporal perception between sensory modalities. The learning of new durations at this stage is better in motor learning (imitation, motor synchronization) than in perceptual learning condition85,86,87. The three-year-old can remember a specific duration associated with an action six months after it was learned87, and already have a good knowledge of the duration of familiar activities88. This leads authors to state that at this age time is still linked to the experience (“experienced time”85) and to the events (“event time”9), and not yet an abstract concept enabling them to voluntarily focus their attention on the time flow. This could also result in an improvement of time estimations from three years of age onward, especially in the temporal generalization tasks that require a comparison between a current duration and a standard duration stored in memory6,89,90,91 (Box 1).

A primary configuration of the core hubs of the DMN, notably medial prefrontal and medial temporal cortices92,93,94,95, together with the lengthening of coherence in an anterior-posterior gradient between four and six years, could support the improvement of interval timing of five89 up to eight seconds90 observed from three to five years old. On the other hand, the incomplete maturation of parietal association cortices and the still-to-be defined boundaries between resting state and task-positive networks, result in limited working memory and attentional abilities, likely linked to the more variable temporal estimates compared to late childhood7,18,27,96,97. This would account for their more variable, fuzzier internal representation of time in memory90,98,99, and their greater time distortion in dual-task and attention distraction paradigms25,100.

During infancy, children are fascinated by the world around them, starting to integrate different aspects of the reality101. However, the cross-modal comparison of speed and time is still difficult to afford. The passage of time feeling thus remains extremely context-dependent, with, for example, the feeling that time passes faster with the increase in the feeling of happiness or when the durations lengthen102 and with ore events meaning more time elapsed103. Indeed, Stojić et al., (2023)104 observed that pre-kindergarteners retrospectively judged one minute of an eventful video lasting longer than an uneventful one of the same length. What is more, when asked to place the silhouettes of three human shapes along a horizontal timeline, three-year-old infants placed the larger silhouettes in the future, while the smaller figures were placed long ago in the past105, reflecting a magnitude effect106,107.

Language development, notably since entering kindergarten, also helps establishing connections between objects and events108, allowing children aged five years old to be able to locate events in the past or in the future9,109. Furthermore, the acquisition of duration words in the development, make four years old children able to comprehend a difference in magnitude between seconds, minutes, hours, weeks, months, and years, even without any knowledge about the precise duration corresponding to these words110. Learning of duration words allowing to denote different moments of the day indeed would make children at this stage already able to understand and to actively be part of their daily routine.

The development of the ability to infer other people’s internal states (Theory of Mind)111,112,113 would enable four years old toddlers to better embody other’s time observed in various situations, thus improving their temporal sensitivity in social contexts114, but also to interact and cooperate with others. From an embodied cognition perspective115,116, the subjective experience of time is grounded to our own mental, emotional, and motor states117,118,119. As an example, from the age of three, children show distortions of time when an angry face is presented120,121,122.

Furthermore, the specialization of the motor pathways, led children to further improve their motor rhythm (e.g., dance, walk and run smoothly). Despite the varying results across studies, the overall of results is that the spontaneous motor tempo is faster (at about 300–400 ms)123 and especially more variable at this age124,125. Also, the ability to synchronize movements to external rhythms (e.g., group dance, singing)126,127 is reflected by a better control in slowing down their pace and a greater tapping regularity from eight years old86. The perception of rhythms and short intervals is fundamental for children at this stage for the beginning of sport activities, as they involve waiting, movements coordination and manipulating moving objects. (e.g., football, baseball, or tennis). Furthermore, time discrimination and number processing is proved to underpin advanced numerical and temporal concepts, as well as aspects of higher order cognition128.

Middle and Late Childhood (6–10 years old) – school years

From a morphological point of view, cortical gray matter volume increase until ten years old, and it is then followed by a linear cortical thinning increasing with age, primarily in parieto-temporal areas129 likely reflecting the synaptic pruning130. At the same time, frontal gray matter was observed to grow with increasing age131. These changes are mirrored at a microstructural level, both by global and regional white matter volume expansion132,133, and increased myelination rate. This augmented speed in axonal conductivity supports an improved information transmission between structurally connected and jointly activated regions, resulting in an enhanced communication between functional networks71,134. At a functional level, EEG studies in healthy five to eight years old children showed that RSNs, such as the DMN, dorso-attentional, cingulo-opercular, ventral and FPN are well formed, but still show an indistinct pattern of activity compared to adolescents135. In late childhood (that is, from six years old to the onset of puberty), the network topology of the brain is reshaped to be more efficient and stronger, with an increase in both local and global efficiency with a specific decrease of connections between adjacent regions (segregation)136. Indeed, although still not completely efficient, there is an improved working memory137, executive control of attention138,139,140 and processing speed141.

The increasing myelination of frontal areas and the white matter volume increase in left precentral gyrus and in left insular cortex142 seen around the age of ten143 and the increasing activation with age of the fronto-striatal and fronto-parietal areas could be pivotal for the improvement of interval timing at this stage144, as a sharp enhanced precision in temporal generalization was reported in eight years old compared to three- and five-years old children90. Improved duration processing abilities were associated to an increased capacity of storage of a reference duration and the comparisons with the probe stimuli18,21,27,97. Moreover, a better duration processing was associated to increased capacities to direct attention to temporal information and to maintain it throughout the entire duration145. The development of these cognitive capacities allows children to better detect the precise start and end of the duration, to better follow and retain the time flow in working memory and to better update it25,100,146. All this, results in better time encoding with less noisy memory representation of time98. Enhanced inhibitory abilities with development also results in lower impulsivity and fewer tendency to press too early in temporal reproduction task and to endure the temporal wait147. Despite a clear improvement compared to toddlers, it is only by nine years of age that children achieve adult-like performances in supra-second duration estimations148. Although time sensitivity is close to the adults’ one18 at 8 to 10 years old with short ( <1 s) and longer intervals ( >8 s), for example, does not mean that there is no improvement from this stage onward. Also, some age differences are still observable between 8-years old children and adults in bisection tasks when a smaller ratio (5:6) is used between two anchor durations ( <1 s and >3 s)97, with a higher variability of performances in time discrimination6. This lower sensitivity has been related to lower selective attentional capacities, or slower processing speed rather than decisional processes, with age being the best predictor of variance of the Weber Ratio in short than long anchor durations. Differences are still evident in temporal generalization paradigms involving short-term memory as suggested by the flattening of the generalization gradient in younger children after the introduction of a retention delay (500 ms, 5 s, 10 s), which was not observed in their adult counterpart149.

These dynamic maturational changes on the structural and functional substrates of temporal processing, concurrently with the wide variety of stimuli and instructions with which the child benefit from during the first school’s years, would boost the development of fluid cognitive skills and expand the vocabulary150,151. Around nine years old indeed, the acquisition of the concept of time allows the child to think about it in several contexts152 and to begin to appropriately use the spatial metaphors of time9,153,154(e.g., to say that the shorter the duration, the faster time passes)102. One can also use time counting strategies to ensure its accurate measurement155 or being to reason about the complex relationship between time, space, and speed156.

In addition, the higher precision in timing abilities reached at this stage is not only crucial for mastering many gross and fine motor skills, but also for estimating the time elapsed since the last event, and the time ahead until the next event. The development of a temporal organization of the events is pivotal to establish connections between past, present and future, therefore structuring our temporal reasoning and logic links, for instance, the concept that physical causes precede effects8. Moreover, the acquisition of calendar’s landmarks, which requires a good representation of the sequence of months over the whole year and a constant update in memory157 contribute to orientate in time and to project themselves into the future, therefore helping the construction of a coherent narration of the self158.

Adolescence (10–19 years old) – high school - college

Adolescent’s brain structure shows higher white matter and lower gray matter volume in the frontal and parietal cortices compared to younger children (i.e., nine years old)159. In prefrontal and parietal cortices, gray matter volume peaks around at age 12160 whereas the one of the temporal lobes continue to expand until 17 years old161. Of note, a steep acceleration in gray matter loss and a considerable white matter increase have been observed in the dorsal prefrontal and orbitofrontal cortex, back over to the posterior parietal regions162. In parallel, the increase of myelination163 is considered to support the improvement of speed of processing documented throughout adolescence164.

At a functional level, the systems projecting from basal ganglia show increased participation of the striatum and precuneus165. Notably, the dorsal striatum, as a part of the thalamo-cortico-striatal circuit deemed to be central in timing166,167 (Box 3) reaches its mature functional connectivity168, starting to establish long-range connections between pre-frontal and the posterior cingulate gyrus, which continue to develop during adulthood169,170. In addition, increasing segregation between control networks, such as the CON, FPN and DMN, has been observed to increase in early adolescence (i.e., 13 years old)171,172,173. The strengthening of communications between basal ganglia and prefrontal corticese may contribute to the improvement of working memory, response inhibition and attentional set shifting174,175, therefore improving cognitive control processes176. Although many of the executive control networks are still in place, the very limited number of studies linking interval timing and brain functional connectivity in adolescence does not allow us to discern which of the executive control processes would benefit the most to temporal precision.

These maturational changes in gray matter within the frontal, parietal177, and striatal159 cortex could support a refinement of the ability to discriminate intervals ranging from 400 to 1100 ms and reproduce durations lasting from two178 to eight seconds5. Indeed, progressive age-related increases in activation in left dorsolateral/inferior prefrontal cortex, and in right hemispheric striato-thalamic and superior parietal regions during the duration discrimination of 500-to-300 ms of difference from a 1000 ms reference interval was observed from 8- to 14-years old children179. Also, by using a temporal bisection task, Li and colleagues (2021)180 distinguished two developmental stages (7–11 and 12–17 years old), suggesting a steady improvement in time sensitivity from the childhood.

However, there is a gap in studies on time perception between nine years and adolescence, probably because duration discrimination abilities at that stage are already close to those of adults148, although differences remain in certain difficult temporal tasks such as the complex kinetic time judgment tasks, fundamental for motor imagery ability181.

Findings of passage of time in adolescent are scarce too31,182, and further longitudinal studies are needed to outline a neurodevelopmental trajectory of this temporal aspect. Awareness of time and its passage, awareness of being subject to time distortions, self-construction, and uncertainty about the future183 should also impact time judgments and the imagination of future scenarios (i.e., mental time travel)184,185,186. Indeed, the imbalance between the development of limbic structures involved in reward and the still immature prefrontal cortical top-down control system187 would lead adolescents to assume more risky behaviors188, therefore underestimating future consequences. On the other hand, the maturation of long-range connections between posterior and anterior parts of the DMN176,189 supports, among other functions190 the ability of “mind-wandering”191,192. The ability to decouple from the present, allows to mentally travel in time, which involves the integration of multiple timescales193. Hence, the mental time travel could be pivotal in this period for the exploration of several identities in order to develop a coherent sense of self.

The adolescence could be considered as a sensitive period during which endogenous (i.e., the brain development) and exogenous factors interact dynamically. Indeed, alcohol and drugs abuse, the exposure to negative life experiences, substance abuse and/or low parental education, increase the vulnerability from adolescents aged 11 to 16 years old onward, to unfold neuropsychiatric conditions, like mood194,195, personality and/or psychotic disorders143,196 which are known to alter the subjective experience of time (for a review2).

Adulthood (19–60 years old)

The maturation during adult age is characterized by a significant acceleration of frontal and striatal gray matter loss188,197,198,199. The caudate and putamen reach 90% of their development after 25 years of age200. The pallidum and the cerebellum instead show an inverted U-shape, starting their gray matter volume shrinking after 25 years old201. On the contrary, myelination of most prefrontal-striatal pathways202,203, notably cortico-spinal tracts, continue later throughout adulthood (i.e., 28 years old)204. Although less markedly, the myelination of thalamus nuclei, caudate and putamen, has been ascertained to increase between 18 and 41 years old205, and to remain stable until sixty years old206.

From a functional standpoint, task positive (comprising dorsolateral prefrontal, precentral, and inferior parietal cortices) and task negative networks (whose hubs lie in posterior cingulate cortex, lateral parietal areas, parts of the medial frontal gyrus and the ventral anterior cingulate cortex)207 become more and more defined by their specific roles, showing the strongest anti-correlations between the right anterior insula, bilateral infero-parietal lobule and the CON at rest176. The increasing anti-correlation between these regions into adulthood, could be an important hallmark of mature executive functions208, linked to sustained attention, working memory and inhibitory capacities209. The gradual development of task-specific fronto-striatal and fronto-parietal networks in the transition from adolescence to young-adulthood was identified as a proper hallmark of brain maturation210,211,212.

Therefore, the enhanced performance to duration tasks both with sub-second98,213,214,215 and supra-second standard intervals26,73,216 observed between eight/nine-years-old children and adults might be associated with the improved connectivity between inferior fronto-striatal-parietal pathways180. Also, the strengthening of the links between right Fronto-Insular Cortex and other hubs of the salience, and executive control network, could be pivotal for integrating information of internal timing signals and external temporal cues, improving timing performances.

Furthermore, improved processing speed, a larger working memory storage and enhanced decision-making abilities in adulthood217 have also been associated with better performances across temporal production, reproduction, and bisection tasks27,218. In temporal generalization task, the asymmetric gradient observed in adults compared to the symmetric gradient observed in children (and also in rats)98,215 has been attributed to changes in decision rules linked to an increase in subjects’ confidence in the precision of their temporal estimation and time knowledge99,219. Moreover, the temporal precision for processing short durations, together with a well-structured temporal cognition organizing days, weeks and months reached during adulthood, make time management optimal at this stage of life.

Further longitudinal studies on every aspect of temporal processing from late childhood throughout adulthood remain necessary to outline the characteristics of knowledge about time and metacognition of time, and their impact at different levels of temporal information processing (e.g., attention, memory, decision)114,219, and the interweaving with several timescales (i.e., minutes and days).

Towards aging (>60 years old)

After the fifth decade of life, the earliest macrostructural changes consist in gray matter atrophy of prefrontal220 superior frontal and insular cortices221,222,223,224. Although non-linearly, gray matter volume changes also occur in subcortical structures, such as hippocampus225 cerebellum226 and striatum227. On the other hand, white matter total volume declines steadily after 60 years old228 and alterations of white matter microstructure follow a posterior to anterior gradient in tracts traversing motor and sensory cortices229 mirroring the rate of demyelination230,231. This has been associated with the slow-down of processing speed of information usually observed in healthy aging232,233. The degradation of structural pathways is deemed to involve a reorganization of the functional connectivity234,235, which becomes more random and less complex236. Indeed, brain networks in aging shows more functional integration and less segregation237, notably between FPN and DMN, leading to less efficiency238. Altogether, these structural and functional changes occurring in healthy aging239,240, notably affecting the fronto-striatal pathways241,242, lead to the impairment of executive control functions243, although differently across individuals244,245.

The observed decrease in functional connectivity between the right fronto-insular cortex and control executive network246 could impact the efficient duration processing on the range of few milliseconds. Increased variability in temporal estimations has indeed been observed in production and reproduction of intervals ranging from 450 to 1750 ms15 and from 480 to 1920 ms247, as well as in bisection tasks248,249, highlighting an even stronger modality effect compared to younger adults250. The disruption of the prefrontal-striatal pathways251 could explain a reduced precision in reproduction of rhythmic sequences of 2 to 3 s interval range252 therefore impacting the sense of nowness253,254, defined as “the amount of temporal stimulation that could be perceived at a time”255. Moreover, under-estimations in aging have also been observed while producing and reproducing longer intervals, ranging from 4 to 14 s and up to 38 s in presence of working memory deficits, slowdown of processing speed256,257,258,259,260 and attentional capacities decreases248,249,261. Furthermore, the error rate of the older adult’s performance increased at the increasing of the complexity of the task262.

On the other hand, changes in episodic memory and future thinking in aging263 could impair the ability to mentally explore other temporal dimensions, and the vividness of the mental time travel reports264,265. In addition, a reduced subjective feeling of temporal distance both for weeks and months has been reported with increasing age266. Although the study of the developmental trajectory of temporal perspectives goes beyond the scopes of this review, subjective temporal distance could be affected by the perceived shortening of their future time perspective267, bearing in mind the idea that time would pass faster now than before the last 5-10 years, or as we get older31,182,268. This finding could be in line with the underestimation of an event duration by representing a smaller time window related to an event duration of several seconds to minutes range in a horizontal timeline269. Altogether, these changes could reflect the hippocampal volume shrinkage270, and disconnections between posterior cingulate and prefrontal cortices271.

In contrast, sensorimotor synchronization abilities, which are considered to rely on primary sensory and motor cortices, supplementary motor area, anterior cerebellum, and basal ganglia, are generally spared with advancing age124,272,273,274. A slower spontaneous motor tempo is nevertheless found in older compared to young adults275. Temporal order judgments were demonstrated to change in aging in both visual276 and auditory277 modalities, this latter being more evident in centenarians278.

The contrasted findings reported by the few studies conducted on the passage of time31,279 make reconstructing a neurodevelopmental framework of this aspect quite challenging. Further longitudinal neuroimaging studies on timing abilities, involving more complex temporal judgments, with changing contexts, are needed to explain how the neural mechanisms supporting several aspects of temporal processing evolve with advancing age. Although in a heterogeneous manner, internal and external factors (e. g., factors contributing to cognitive reserve, social stress, social isolation, cognitive load) could contribute to the different trajectories of age-related neurophysiological changes in temporal perception and cognition. This interplay could explain the larger variability observed in duration processing compared to adults. Distortions of temporal skills could impact on older individuals’ ability to respect a scheduled appointment or treatments, by forgetting to take the medications at a planned hour. Timing difficulties could not only affect the quality of life of seniors, but also, for example, the ability to manage weekly appointments or to cross the street, thus restricting their autonomy in the daily life. In addition, aging at work, the feeling of not being able to manage one’s own time, for example, can increase stress, which in turn, impact judgments of time280. Numerous studies have also shown the role of emotions and their regulation on time judgments281. Conversely, having a whole and rich collection of personal memories forming our own autobiographical memory, help to maintain a stable self-identity, avoiding disorientation in aging. Deficits of time processing performances in aging might also serve as behavioral marker for pre-clinical stages of dementia282(Box 4).

Concluding remarks

Throughout the lifespan, we face the constant challenge of providing adapted behaviors in complex and changing environments. Such adaptation requires temporally specific representations and actions across time scales and temporal judgments1. This ability is constrained by neurobiological maturation and evolution, in response to specific environmental challenges. Here, we reviewed the milestones of the maturation of neural and cognitive underpinnings of time processing form gestational period to senescence in the attempt to conceive a neurodevelopmental model of the emergence of each aspect of time processing (order and rhythmicity, duration processing, passage of time and mental time travel; for a review see ref. 2). The efficient integration of these aspects in every day’s life supports a wide array of survival functions, such as motor control282, action coordination283 and language284,285. An integer sense of time support consciousness’ mechanisms286, fundamental for one’s own grounding in the moment282 as well as for temporal anticipation of future events285,287.

Current models of time perception, being based on central or distributed timing mechanisms, only took in account timing mechanisms underlying the processing of short durations (from milliseconds to few minutes) in healthy young subjects (Box 2), ignoring their developmental trajectory. We encourage future investigations to specify how the current models can account for data acquired across age groups. Here we propose an additive maturational model (Fig. 2), based on functional specialization, that could be named RDPM as Rhythm (Simultaneity and Temporal Order) – Durations - Passage Of Time - Mental time travel. That is, these four temporal aspects would emerge following this order. The emergence of each aspect will add to the previous one, enriching temporal cognition, due to the interaction between “internal factors” (i.e., structural and functional specialization of targeted brain areas, neurotransmission and neuroendocrine regulation) in the attempt to cope with the environmental challenges, namely “external factors”. According to this framework, the first sub-cortical structures to set up, such as basal ganglia, cerebellum and primary sensory cortex promote the development of rhythmicity already before birth, in the attempt to synchronize with mother’s movements to establish a sort of interaction (R). Then, newborns and infants began to handle short durations (D) by the mean of myelination increase and cortical volume growth of frontal areas. Nevertheless, it is only from middle to late childhood that the processing of longer durations became akin to the adult one, thanks to the maturation and specialization of higher-order cortices and the myelination of frontal-striatal pathway promoting top-down attention regulation and working memory. Growing up, the feeling of the passage of time (P) appears and refines itself concurrently with the development of parietal and inferior and medial temporal cortices, by the need to integrate speed, space and duration. At last, Mental time travel (M) ability would only appear thanks to the refinement of prefrontal cortices and the flourishing of connections between temporal and pre-frontal lobes in adolescence, to support abstract reasoning since adolescence. The neuroanatomy of a temporal circuitry would set up progressively, involving at first only subcortical areas, to then extend itself to higher-order cortices (Fig. 2).

Overall, the evolution of temporal abilities throughout lifespan seem to follow the “last in, first out” hypothesis, according to which the most recent brain areas to be developed are the first to be affected in aging287,288. Therefore, the most primitive and the first timing capacities to be developed such as rhythmic processing289,290,291 are the ones that last longer, while the most sophisticated and the last to appear mental time travel is the first to change in aging.

Nevertheless, further longitudinal neuroimaging studies are needed to elucidate the neurodevelopment of the passage of time judgments and their relationships to duration judgments throughout the life span to reconstruct a complete neurodevelopmental progression of the temporal processing (Box 5). Specifying a healthy trajectory of the development of brain and cognitive substrates supporting timing abilities, could expand our understanding of neural mechanisms and cognitive models of temporal processing, shedding light on the intricate debate among distributed and centralized models (Box 2).

To conclude this lifespan approach could clarify the evolution of temporal mechanisms underlying the Rhythm (Simultaneity and Temporal Order) (R), Durations processing (D), Passage of Time (P) and Mental time travel (M). Also, this perspective would foster future longitudinal works to assess the neural bases and cognitive functions linked to each of these temporal aspects. Furthermore, we encourage to consider alterations of temporal cognition and its interweaving with neuropsychiatric conditions which may unfold throughout the life-course in clinical settings292,293. In this way, it would be possible to promote the design and the clinical use of cognitive batteries assessing temporal performances across temporal mechanisms. Evaluating temporal perception and cognition could therefore be useful for the detection of neuropsychiatric disorders of the development and neurodegenerative disorders early in the latest stages of life, improving the development of new treatments and disease management (Box 3).