Elsevier

Cognitive Brain Research

Volume 21, Issue 2, October 2004, Pages 193-205
Cognitive Brain Research

Research report
Neural representation of interval encoding and decision making

https://doi.org/10.1016/j.cogbrainres.2004.01.010Get rights and content

Abstract

Our perception of time depends on multiple psychological processes that allow us to anticipate events. In this study, we used event-related functional magnetic resonance imaging (fMRI) to differentiate neural systems involved in formulating representations of time from processes associated with making decisions about their duration. A time perception task consisting of two randomly presented standard intervals was used to ensure that intervals were encoded on each trial and to enhance memory requirements. During the encoding phase of a trial, activation was observed in the right caudate nucleus, right inferior parietal cortex and left cerebellum. Activation in these regions correlated with timing sensitivity (coefficient of variation). In contrast, encoding-related activity in the right parahippocampus and hippocampus correlated with the bisection point and right precuneus activation was associated with a measure of memory distortion. Decision processes were studied by examining brain activation during the decision phase of a trial that was associated with the difficulty of interval discriminations. Activation in the right parahippocampus was greater for easier than harder discriminations. In contrast, activation was greater for harder than easier discriminations in systems involved in working memory (left middle-frontal and parietal cortex) and auditory rehearsal (left inferior-frontal and superior-temporal cortex). Activity in the auditory rehearsal network correlated with memory distortion. Our results support the independence of systems that mediate interval encoding and decision processes. The results also suggest that distortions in memory for time may be due to strategic processing in cortical systems involved in either encoding or rehearsal.

Introduction

Interest in how the brain processes temporal information has grown over the years, due to its importance in everyday activities that depend upon anticipating events and flexibly adjusting behavior to changing temporal goals. Information processing theories of temporal cognition maintain that multiple processes determine our ability to time events [20], [33], including a metaphorical “clock” that represents time through the accumulation of pulses emitted from a timekeeper mechanism. The level of attention devoted to the passage of time is thought to influence the operation of the timekeeper, perhaps by mediating the starting and stopping of pulses. Once a representation of time is formulated, it is encoded into memory, and then decision processes compare the current representation of time from the clock process with those stored in memory to decide when and how to respond. The interaction among these component processes gives rise to our perception of time.

The present study investigated the neural representation of processes associated with encoding intervals and making decisions about their durations. The neural underpinnings of temporal information processing remain controversial, largely due to the difficulty in distinguishing between systems that specifically support timekeeping mechanisms and systems that regulate other processes, such as decision making. Focal lesion studies in humans have shown that damage to the basal ganglia, cerebellum and cerebral cortex all disrupt the ability to perceive and/or reproduce time intervals or perform other behaviors that appear to depend upon accurate timing [8], [21], [22], [23], [25], [35], [39]. The reasons for performance impairments remain controversial, because it is unclear how a deficient timekeeper should affect temporal processing proficiency (e.g., accuracy, variability) relative to impairments in other processes such as attention [55] or motor dysfunction in the performing limb [18]. Adding to this problem, functional imaging research has focused on motor timing or used blocked-trial designs, which make it difficult to distinguish the role of central timing mechanisms from sensorimotor or other processes.

In an earlier study, we addressed some of these limitations by using the temporal resolution capabilities of functional magnetic resonance imaging (fMRI) to study activation in neural systems associated with different components of a time perception task [45]. In this task, two tones separated by 1200 ms (defining the standard interval) were presented, followed by a comparison interval. Subjects judged whether the comparison interval was longer or shorter than the standard. We showed that activation in the basal ganglia (bilateral caudate and putamen) and right inferior parietal cortex developed in association with encoding the standard interval. This contrasted with activation in the right dorsolateral prefrontal cortex (DLPFC), which developed later in association with comparing the two intervals and making a decision about their relative duration. These results were consistent with the respective roles of the basal ganglia and the right inferior parietal cortex in timekeeping [36] and attention [23] components of the clock process and the right DLPFC in executive components of working memory [43].

Despite converging evidence supporting the striatum and dopamine neurotransmission in hypothetical clock processes [22], [32], [36], [38], [39], other research suggests that the caudate nucleus is intimately involved in working memory [29], [42]. This proposal suggests that early caudate activation in our study could be due to actively maintaining the same representation of a single standard interval across trials, rather than encoding the interval on each trial. We addressed this issue in the present study by randomly presenting two different standard intervals (1200 or 1800 ms) to encourage subjects to encode the standard on each trial. If the caudate is involved in formulating representations of time, we predicted that activation should be seen within 4 s after the onset of the standard interval.

A second focus of the study was to investigate the neural systems that support decisions about two temporal events. Decision processes are involved in comparing pulse counts from the clock process with those stored in memory, but little is known about the nature of these processes. Theoretically, decisions take into account a threshold for determining whether a comparison interval exceeds the duration of a standard interval [20]. This assumption is consistent with a study showing that decision thresholds can be biased by manipulating payoff contingencies for detecting correct or incorrect responses [56]. In this study, selective attention to reinforced responses affected decisions about temporal events, but not their timing, implying a functional independence of the clock and decision components of timing. Selective attention has been implicated in making decisions about time in electrophysiological studies [9], [10]. Here, the amplitude of slow cortical potentials in the prefrontal cortex is greater for incorrect than correct responses, ostensibly reflecting the lower level of attention paid to time when intervals can be easily discriminated. Still, the precise source(s) of the electrophysiological responses is (are) unknown, as is the role of other neural systems. Additionally, electrophysiological responses distinguishing correct and incorrect trials could also reflect the quality of interval encoding, rather than decision processes per se. Complicating the identification of decision processes is the fact that decisions closely follow the processing of information that must be acted upon. It is therefore not clear whether neural activity associated with selecting a response is due to the goodness of encoding an interval, decision making, or both. For these reasons, we examined the effect of time discrimination difficulty on brain activation during the decision-making phase of a trial to evaluate regions involved in decisions more directly. This aspect of the study also allowed us to test the independence of neural systems that support clock and decision components of temporal information processing. We predicted that decision difficulty should not influence activation in systems that are principally involved in timekeeping operations, if these processes are independent [20], [56].

A related issue pertains to the role of memory in interval encoding and decision processes. Timing theory assumes that output from the clock process is encoded into memory and then retrieved for decision making, suggesting that the medial temporal lobes (MTL) should participate in temporal processing, given their role in memory [7]. It has been speculated that the MTL is crucial for keeping memory traces active and accessing them for decision making [7], [47]. However, controversy remains as to whether MTL lesions in animals disrupt timing [13], [40]. Functional imaging studies have not found MTL activation during timing, although this could be due to the use of “control” conditions that subtract out activation. Another explanation relates to the use of a single standard interval across blocks of trials, which could minimize memory demands during temporal processing. This possibility is suggested by the “migration” of temporal estimates when subjects are trained on two different intervals [30]. When different intervals are tested together, shorter intervals are overestimated and longer ones underestimated relative to when they are tested separately. This appears to reflect a mixing of memories for the two intervals rather than a decision bias, because the latter should produce a distortion in the same direction for both standard interval conditions, which does not occur. In the present study, we expected that the use of two randomly presented standard intervals would place greater demands on memory processes, resulting in MTL activity in association with interval encoding and decision making.

Finally, while we assumed that clock and decision-making processes were operating primarily during the standard interval encoding and decision phases of a trial, respectively, other processes are likely ongoing at the same time during both phases. To better define the functional significance of activity during these two phases of a trial, we correlated brain activation with different behavioral measures of time perception to better identify systems that were preeminent in interval timing. Whether patterns of brain activation during temporal processing can be distinguished by measures that have different theoretical significance is not known. Although the bisection point, a measure of accuracy, reflects clock speed [20], we did not expect it to correlate with timekeeping systems in our study since the rate of the clock should be the same for both standard intervals. The point of bisection might correlate, however, with activity in systems that encode and retrieve representations of different intervals. We expected the difference between the bisection points, a measure of the migration or distortion in memory for intervals [30], to correlate with activation in systems associated with memory processes. Finally, we predicted that the coefficient of variation, a measure of temporal processing efficiency or sensitivity [20], would correlate with activity in systems previously associated with timing in lesion studies including the basal ganglia [22], [39], right middle-frontal and inferior parietal cortex [23] and cerebellum [25].

Section snippets

Subjects

Study participants were 24 right-handed healthy adults between the ages of 21 and 53 (mean=30.6, S.D.=10) with no history of neurological or psychiatric disorders. Only nonsmokers were studied due to the effect of nicotine on the dopamine system. Participants were asked to refrain from drinking alcohol 24 h prior to their scanning session. Study procedures were approved by the Human Research and Review Committee at the University of New Mexico Health Sciences Center. Informed consent was

Behavioral results

A repeated-measures ANOVA tested the main effects of standard interval, comparison interval and their interaction for RT and percent longer responses. The Huynh-Feldt correction was used to adjust for heterogeneity of variance. In the analysis of the RT data, there was a main effect of comparison interval [F(5.3,161)=3.49, p<0.01], showing that RTs were generally longer for comparison intervals closer than further away in physical time from the standard interval. RTs in the 1200-ms condition

Interval encoding

The results showed that distributed regions of the basal ganglia, cerebellum and cerebral cortex were activated during the standard interval encoding phase of a trial. However, activation correlated with behavioral measures of time discrimination in only some of these regions including the right caudate, right inferior parietal cortex and precuneus, right parahippocampus and hippocampus and left cerebellum. Damage to most of these regions produces timing deficits [22], [23], [25], [40], which

Acknowledgements

We thank Sally Durgerian, Alison Lindsay, Kim Paulson, Corby Dale, Gabrielle Mallory, Jennifer Hogan and Ting Lee for their research assistance. This research was supported by grants from the MIND Institute, the Department of Veteran's Affairs and P01MH51358.

References (57)

  • B.R. Postle et al.

    Dissociation of human caudate nucleus activity in spatial and nonspatial working memory: an event-related fMRI study

    Cogn. Brain Res

    (1999)
  • O. Simon et al.

    Topographical layout of hand, eye, calculation, and language-related areas in the human parietal lobe

    Neuron

    (2002)
  • E. Tulving et al.

    Memory beyond the hippocampus

    Curr. Opin. Neurobiol

    (1997)
  • J.H. Wearden et al.

    Manipulating decision processes in the human scalar timing system

    Behav. Processes

    (2003)
  • G.I. Allen et al.

    Cerebrocerebellar communication systems

    Physiol. Rev

    (1974)
  • S.J. Blakemore et al.

    The cerebellum is involved in predicting the sensory consequences of action

    NeuroReport

    (2001)
  • J.W. Breukelaar et al.

    Effects of lesions to the cerebellar vermis and hemispheres on timing and counting in rats

    Behav. Neurosci

    (1999)
  • J.B. Brewer et al.

    Making memories: brain activity that predicts how well visual experience will be remembered

    Science

    (1998)
  • M.A. Burock et al.

    Randomized event-related experimental designs allow for extremely rapid presentation rates using functional MRI

    NeuroReport

    (1998)
  • L. Casini et al.

    Effects of divided attention on temporal processing in patients with lesions of the cerebellum or frontal lobe

    Neuropsychology

    (1999)
  • L. Casini et al.

    Relation between level of prefrontal activity and subject's performance

    J. Psychophysiol

    (1999)
  • J.E. Desmond et al.

    Lobular patterns of cerebellar activation in verbal working-memory and finger-tapping tasks as revealed by functional MRI

    J. Cogn. Neurosci

    (1997)
  • A. Dietrich et al.

    Is the hippocampus involved in temporal discrimination and the memory of short intervals?

    Int. J. Neurosci

    (1997)
  • W. Fias et al.

    Parietal representation of symbolic and nonsymbolic magnitude

    J. Cogn. Neurosci

    (2003)
  • P.C. Fletcher et al.

    Frontal lobes and human memory: insights from functional neuroimaging

    Brain

    (2001)
  • P.C. Fletcher et al.

    Brain activity during memory retrieval. The influence of imagery and semantic cueing

    Brain

    (1996)
  • S.D. Forman et al.

    Improved assessment of significant activation in functional magnetic resonance imaging (fMRI): use of a cluster-size threshold

    Magn. Reson. Med

    (1995)
  • E.A. Franz et al.

    Reduced timing variability in patients with unilateral cerebellar lesions during bimanual movements

    J. Cogn. Neurosci

    (1996)
  • Cited by (0)

    View full text