Neurobiology of interval timing published: 10^th July 2025

What, then, is time? If no one asks of me, I know; if I wish to explain to him who asks, I know not.
- Saint Augustine

For centuries, humans have thought about time, its nature and how we process times in our minds. Time perception is part of the human experience yet the complete neural basis of time perception is still unknown. One of the most fundamental variables that organisms have to quantify in order to survive is temporality or time. Our brains have to develop a sense of time to do activities like talk, sing, run, perform activities in a group, drive, etc. Humans unfortunately do not posses a sensory organ to detect time like we do for audio and visual sensory input yet we have a sense of time. Time thus is not a material object of the world for which we could have a sensory organ to sense it. On a scale from microseconds on one side and circadian rhythm on the other, temporal information guides our behaviour. A 12 order magnitude is covered by this scale.

Microsecond Scale

Information processing in the microseconds is needed for binaural hearing and echolocation. For binaural hearing the microsecond scale is used by the auditory system to determine the differences in time in the action potentials for the two ears called as interaural delay to determine the spatial origin of the sound. The sound will arrive slightly earlier for the near ear usually in the order of 600-700μs because of the natural interaural time difference (ITD).

Milliseconds Range

Intercal timing ranging from 200-1000ms involves a broad spectrum of activities ranging from collision avoidance, object interception, speech perception, motion processing, execution and processing music, coordinating fine movements, etc. The temporal processing in this range is quite sophisticated yet poorly understood on how the neural processing works to make it happen.

Seconds to Minute Scale

Temporal processing in this range depends on conscious and cognitive control and is involved in foraging, decision making, sequential motor performance, arithmetic, associative learning, etc. Timing in this scale forms the contextual framework through which our behaviour is mapped onto the external world.

Circadian Rhythms

The biological timing that organises the daily environmental oscillations every 24 hours is known as the circadian clock. The two major functions of this clock are:
- Prepare your body for what's coming next based on the time of day. It optimises the temporal manifestations of different biological activities during the day by anticipating of recurring fluctuations in the environment.
- Prevent biological conflicts by timing different processes to not interfere with each other. Eg. release growth hormones only while sleeping and not during the day when you need energy for activities.
The circadian rhythm in mammals is governed by a complex network of cellular molecular oscillators distributed throughout the brain and peripheral tissues. The master clock is in the brain in this region known as hypothalamic suprachiasmatic nuclei (SCN).

It synchronises the internal clock with the external light-dark cycle of the day. The internal circadian clock in mammals possesses a rhythm with an approximate 24 hour period and the synchronising signal is light. The SCN acts as a relay between the external light-dark cycle and the internal timing system. So, the peripheral clocks are governed directly by the SCN that controls the activities like feeding, body temperature, hormone release, etc.

Different timing mechanism for different timing behaviours?

The word timing can have the connotation of either how long an event lasts or when an event occurs. So, the internal clocks should have the ability to encode the elapsed time from a stimulus or process such as the time between the notes in a song, along with the capacity to measure remaining time for an action where the system should select a precise moment for doing something for an accurate result like when a batsman hits a cricket ball.
Our brains need to compute time from different modalities of sensory stimulus or when involved in rhythmic activities like music and dancing or when marking the elapsed time from an event or as the time remaining for an action. So, some key elements of temporal processing include the time scale being quantified, the modality of the stimuli guiding the time perception, whether the time is being measured for a movement or perceptual decision, whether the task involves single or multiple intervals, whether the timing generated is external or internal.

Cognitive model of time

Research from experimental psychology, clinical neuropsychology, and neuroimaging have produced an extensive literature on the cognitive mechanism of temporal processing. Cognitive models distinguish time perception with two fundamental views: prospective and retrospective.
The prospective model says that an observer judges the duration of an interval that is presently being experienced. This model assumes of an "internal clock" with a pacemaker producing a sequence of time units that are fed into an accumulator. Here the time units produced are only registered when attention is directed to time. This makes prospective model dual in nature as the observer has to switch focus between temporal and non-temporal processes while the accumulator collects the time units generated by the pacemaker in the background. This model is also called the Scalar Expectancy Theory. An internal clock, working memory, a reference memory, and a decision process is required as per this model. The onset of a stimulus causes the switch to connecting the pacemaker and the accumulator to close allowing the pulses or the ticks of the inner clock to flow. The offset of the stimulus causes the switch to open, cutting the connection. The resulting pulses represent the elapsed time and is sent to the working memory. While timing an event the value in the working memory i.e. the current time is compared to the value in the long term reference memory i.e. the expected time. When the two values are close an action is initiated; when they are far apart the action is stopped. To make the comparison, the ratio of the two values are computed; when the ratio is less than a certain value an action is initiated else no action is initiated. The key point here is that a ratio of the two values is compared instead of subtracting the absolute values i.e. timing precision is relative to the size of the interval being timed. Hence the name "Scalar". For eg. when timing a 10 second interval the observed value might be precise upto 1 sec but if timing a 100 second interval the precision might be about 10 seconds. However, a cyclical pattern can be observed here. You need a reference memory to compare the durations and these references are created from previous duration experiences which themselves require duration judgments with references that do not exist yet. How do you compare your first duration when you have no stored pulse counts? The model works well for explaining timing behaviour but struggles with the foundation question of how temporal quality emerges initially. Clocks, were developed on the original capacity we have to subjectively experience time; they are a product of our mental, physical, and psychological capabilities. We created the devices like clocks, chronometers, watches, etc. because we could already perceive duration, sequence, and a temporal flow. Clocks produce a measurable time for us which are a result of having a temporal experience and we developed them to explain the same experience; it is circular to explain how time works by pointing to a clock.
The retrospective model of temporal processing says, the observer has to estimate the already elapsed time interval duration. The observer has to retrospectively estimate a given duration from the amount of processed and stored memory contents i.e the duration has to be constructed from memory. This model explains why certain events feel longer in duration than others. Our experience of time is strongly influenced by cognitive factors such as the amount of information processed during the interval, the complexity of the stimuli, the way the stimuli was stored in memory, memory load, level of attention, etc. That is to say the more changing experiences we have during a certain timespan the longer the duration is subjectively experienced. So a routine activity if compared to a new or novel activity leads to the perception of shorter duration. The judgments that we make in retrospective about the duration are based on intervals spanning a few seconds (short term memory) to a whole life time (long term memory). A storage size metaphor is applicable here to explain this model. Duration as a mental construction from memory is formed from the size of the information stored in a given interval. It is the information in the storage that determines duration experience. However, just like the prospective model this model too does not account for all the other modes of experiencing time. In fact it was shown through experiments that if observers are asked to prospectively judge a time interval and perform a complex task concurrently during the interval, their perception of the interval duration shortens instead of extending as a function of the amount of non durational information processed in the interval. Another factor that strongly influences retrospective duration estimate is contextual changes in the form of environmental, psychological, interoceptive, etc. that makes the observer think that the judged interval was longer in duration which the storage size model cannot account for. The reliance of the retrospective model on the human memory leads to flaws in itself. Human memory does not work like computer memory. Information stored as memory in the brain is constantly being shuffled, reinterpreted, and interconnected with other memories; the information does not stay static. The retrospective model assumes we have discrete information units stored in our brains like computer files and we use them to judge durations. When you recall something from your memory you are not just recollecting the memory, you are also reconstructing the experience.
As you can see, no matter what model we come up with to explain the temporal experience it always falls short to account for the full experience of time and keeps us looking for more ways to understand how exactly the brain processes time.

Neural model of time

Currently we have two opposing models how these different timing stimuli are processed in the brain. The hypothesis of the first model is that there is a common mechanism that processes temporal information across all the behaviours. This hypothesis is supported by classical psychophysical (a branch of psychology that studies relation between physical stimuli and mental phenomena) observations eg. perception and motor production share common timing mechanisms and by brain lesion and functional imaging studies. On the other hand the hypothesis from the opposing model says the temporal processing is ubiquitous and localised depending on the behavioural context. It suggests that time processing does not have a dedicated parts in the brain but rather a property that emerges from the normal functioning of whatever brain networks are involved in the task. This view is supported by modeling, brain slice recordings, and new psychophysical approaches. There is an emerging third view which suggests the existence of partially distributed timing mechanism, integrated by main core structures of the cortico-thalamic-basal ganglia circuit (CBGT) and the areas of the brain that are selectively engaged in the behavioural task and is supported via recent psychophysical studies and functional imaging meta analysis.
Studies done in retrospective model of timing have yielded the hypotheses that the fronto striatal circuits loop between the supplementary motor area (SMA), caudate putamen and pallidum which form a part of basal ganglia are modulated by dopamine systems and are critical for processing interval duration. Individuals with damage to frontal lobes or injuries predominantly affecting the frontal areas show impaired estimates of temporal intervals. In fact, various neuroimaging studies show how dopaminergic systems affect our time perception. Haloperidol is an antipsychotic medication that blocks dopamine D2 receptors in the brain and it impairs the duration discrimination abilities in healthy individuals. Studies in both humans and animals have shown that dopamine agonists like methamphetamine and antagonists like haloperidol influence the timing process presumably by increasing and decreasing the clock speed respectively. Individuals dependent on cocaine and methamphetamine who have structural changes in the dopaminergic targets such as the frontal cortex show impaired time processing on various timing tasks. Hence, an intact dopamine neuro-transmission within the frontal areas of the brain is a crucial requirement for accurate temporal processing abilities.

Sensory Timing

All organisms extract temporal information from stimuli of all sensory channels even if there is no time sensory organ. We are still not sure how time is computed from the activation of different sensory systems nor where is the temporal information extracted in this hierarchy of sensory systems. But let's first understand the the general functional arrangement of the auditory, visual and somatosensory (relating to a sensation such as pressure, pain, or warmth which can happen anywhere in the body) systems that are involved in the temporal processing particularly in the hundreds of milliseconds range. These sensory systems have the following commonalities: the sensory trigger of a signal of the physical information into action potentials in the sensory receptor; the projection of this information through the thalamus to the primary sensory areas of the cerebral cortex; the processing of the different aspects of the stimuli in the cortical circuits of the brain; and finally the use of high order sensory processing for perception, learning, memory and voluntary motor action. Initially, time information could be extracted from the trigger of the signal of the sensory input in the first relays of the sensory systems involved. Not just that various studies show that auditory modality has the strongest ability to extract temporal information in the range of tens of milliseconds across the first relays of the sensory processing which tells us that time is a fundamental behavioural parameter for audition.
For vision, duration tuned cells show up in the first nodes of the visual pathway. Some of the neurons respond best to stimuli that last a specific amount of time. The time range is between 50 to 400 milliseconds. Interestingly, no such tuned cells were found in thr thalamus for visual processing which suggests that time processing in visual stimuli is localised in the first relays of neurons. However, research has shown that temporal processing observed psychophysical tasks in humans is more precise and accurate when the intervals are defined by auditory than visual or tactile stimuli.
Another important property of temporal processing is that the perceived time is affected by whether the intervals in the stimuli are empty or filled. For auditory input empty stimuli could look like: click - silence - click. For visual stimuli it could look like flicker - darkness - flicker. The filled stimuli for auditory stimulus could be a stream of noise or tone and for visual it could be just a flash for the whole interval. Studies have shown that intervals with filled stimuli feel longer than the stimuli with empty intervals. No single mechanism fully explains the effect, it is the result of fundamental differences in how the brain processes temporal information when additional sensory processing is required versus when timing can operate in isolation.
The temporal processing hence seems to be dependent on a specialised group of cells in the early nodes of the sensory processing that are tuned to the duration of auditory or visual stimuli. Also, the multimodal temporal processing of tens of hundreds of milliseconds is anatomy dependent (eyes vs ears) and audition clearly has an advantage in the perception of passage of time.

Time Perception

The integration of duration information across all the sense seems to be dependent on the exstratriate cortex of the brain which is the region next to the occipital cortex which processes visual sensory information. The recognition and perception of passage of time across the senses can be used for planning a motor action, discriminating between intervals of time, categorisation tasks, etc. Multiple studies have been do to see how the brain performs interval learning and discrimination on multiple sensory modalities like audition and somatosensory stimuli. Subjects were trained to recognise a 70-200ms interval between stimuli and it was observed that intensive learning can generalise across modalities, skin location, and brain hemisphere. In one of the studies the subjects were trained on a base interval of 75ms and a target interval of 125ms and the discrimination of the intervals improved significantly on training of the two intervals. In fact interval training for somatosensory stimuli on one hand was generalised for the same interval on the other untrained hand. For auditory stimuli the learning transfer was dependent on improved temporal processing and not on more efficient memory or decision process. These findings not only support the view of a centralised or a partially overlapping distributed timing mechanism but also the existence of duration specific circuitry for temporal processing. Because of the learning the timing signals from the early cortical areas of the brain are sent to the core timing network of the brain which improves the efficiency of processing temporal information. fMRI studies show an early and late visual and auditory areas are activated during the perception of interval tasks. Overall, we see the idea emerging that temporal processing is a complex phenomena which involves cortical and subcortical structures of the brain.

Motor Timing

Time interval processing in the range of milliseconds is required for various complex activities like speech perception and production, perception and execution of music and dance, performance of sports, etc. Time in music comes in different sequences and of different durations; regular interval of different durations on a pulse forms a meter. The ability to interpret the beat and rhythm helps us appreciate music and be able to dance to it. Music and dance are complex behaviour dependent on processing intricate loops of timing information from the sensory organs or generating internally timed motor sequences in case of dancing. Various fMRI studies have shown activation in cortico-thalamic-basal ganglia circuit when involved in time perception activities which are the same areas that get activated in planning and executing motor activities. This region of the brain also includes medial premotor areas (premotor cortex) and presupplementary motor areas (motor cortex). These studies provide an evidence for CBGT being a core element of the timing network for interval processing from the provided stimuli however, there seems to be no evidence for its involvement in capturing the temporal resolution during temporal processing.
Studies show the involvement of medial premotor areas and the neostriatum (a cluster of nuclei in subcortical basal ganglia) for duration tuning on a set of tapping tasks done with monkeys. Such studies again confirm the existence of interval tuning in the core timing network of the brain. Interval tuning here means that there are a set of neurons in the CBGT tuned to a specific interval (let's say 100ms) where they fire with the highest intensity. For different tapping intervals and modalities (visual cue, auditory cue) the neurons showed an affinity for their tuned interval across all the tasks. For instance a neuron tuned to 200ms stayed tuned to 200ms whether the monkey was
- tapping to visual cues
- tapping to auditory cues
- tapping internally without a cue
- making single or multiple taps
This confirms the existence of tuned neurons which are context independent for temporal processing in the CBGT. It's like having an universal stopwatch for 200ms detection. In contrast, the neurons in the lower level sensory processing (auditory and visual) are modality specific, which means visual neurons only respond to visual stimuli and auditory neurons only respond to auditory stimuli and hence localised temporal resolution and processing.
In various experimental psychology studies it was observed that the temporal processing improved significantly when there are repetitions of the standard interval presented to the subjects. Just Noticeable Difference is the amount some aspect of a stimulus must be changed in order for a difference to be noticeable, detectable at least half the time between two stimuli. The standard interval is the interval chosen for various tasks in the studies mentioned above. Multiple exposure to the standard interval improved the accuracy of the subjects whether they were performing visual or auditory tasks for time interval perception or temporalised tapping. Recordings of the cells in premotor areas of monkeys producing multiple intervals in a sequence showed temporal improvement possibly due to an increase in the perceived or executed intervals. The interval tunes cells in the CBGT showed a multiplicative scaling i.e a proportional increase in their response for more produced intervals. The decrease in the variance of the temporal information could be because of the discharge rate of the interval tuned cells. The discharge rate is the rate at which the neurons fire and is measured in spiker per seconds or hertz (Hz). Higher the discharge rate higher the neuron signal which leads to better noise to signal ratio improving the overall accuracy.
One theory of why the neural circuitry of timing shares so much with motor function is that our general sense of time has been developed through actions since childhood. Developmental studies have shown that young kids appear to represent time in motor terms; for them time is how long does it take to do something. They are able to accurately determine an interval if that interval has a motor action rather than being empty. They also find it difficult to dissociate a duration from the motor act itself. A young kid might think a task that required more physical effort took longer even if the durations were identical. It is possible that in case of temporal processing, the learned associations between particular actions and their durations have been engrained in the motor networks. Timing is crucial to perform any motor activity. What is the point of swinging your bat at the wrong time when the ball has already hit the stumps?

The human ability to perceive and generate precise timed intervals is most evident in musical perception and performance. Studies of brain potentials have indicated that human brain has this innate ability to predict rhythmic sensory events and error signals emerge when the music fails to mee the metric expectations. These error signals are measurable electrical patterns in the brain that indicate a mismatch between what we predicted would happen and what actually happened. Our sense of musical time is not passive; we are actively monitoring and predicting the temporal structure of music. It is evident that appreciating music does not just activate auditory cortex but it also engages our motor cortex to follow the rhythmic pattern engrained in the music. The joy of music comes from this ability to predict such rhythmic patterns. As we have seen the ability to predict sensory events and generate precisely timed actions seems not to only depend on the localised general purpose timing circuit, rather every neural network engaged in the sensory perception incorporate time as an essential feature of the information it is processing.