Related articles

• Brain tells time by morphing
• High-fat diet can mess with body clock
• Babies' internal clock development hindered by artificial lights
• Manic depression may be a rhythm thing
• Einstein is still right on time

The circadian clock controls when we feel tired and when we feel alert, but its influence extends much further. It governs daily cycles of body temperature, blood pressure, hormone secretion, metabolism and gene activity, with cycles expressed in almost every cell in the body.

Decades of research have given chronobiologists a fairly good understanding of the cogs and gears that drive the circadian clock. Two clusters of 10,000 nerve cells, one in each hemisphere of the brain, form what Dunlap refers to as the clock's "central pacemaking tissue". These suprachiasmatic nuclei (SCN) sit a couple of centimetres behind the bridge of the nose, in a brain region called the hypothalamus.

Light isn't needed for the SCN to establish a cycle, as the experiments of Siffre and others have shown. But without daylight, our internal timepiece lags behind the 24-hour succession of day and night. Normally, to calibrate the clock to the environment, dedicated cells in the retina of the eye transmit information about light levels to the SCN.

But as Dunlap explains, our circadian cycles are deeply hardwired, so trying to adjust the time by more than a few minutes a day – in a flight from Sydney to San Francisco, for example – can cause debilitating jetlag. "The body has an enormous amount of temporal inertia," he says. "It's like trying to turn around an aircraft carrier."

So if the body has such a powerful and pervasive internal clock, why did Siffre lose his grip on time? "I feel motionless, but at the same time I feel as though I am being pulled along by the uninterrupted flow of time," he wrote. "I try to grab hold of it but...realise I have failed."

According to researchers, while Siffre's body clock was ticking steadily along, his conscious perception of time was being governed by a much more malleable timepiece.

BIOLOGISTS TRADITIONALLY DIVIDE human timekeeping abilities into three domains. At one end of the scale is the circadian clock that set up Siffre's predictable rhythm.

At the other end is the millisecond timing involved in fine motor tasks and speech. Both of these systems are fairly well understood and fairly inflexible. Not so the third domain – the minutes-to-seconds range known as interval timing – which is responsible for our conscious perception of time.

"The mind is a time machine," says Catalin Buhusi of the Medical University of South Carolina in Charleston, USA. He studies the interval timing system – the system that warps time to make it fly when we're having fun and crawl along when we're bored. But while the feelings associated with this flexible system might be familiar, its biological basis remains something of a mystery.

According to Buhusi, the conventional understanding of interval timing is based on a 'pacemaker-accumulator' model. This proposes that the brain has an internal pacemaker that emits regular pulses, a bit like a ticking clock. These pulses are temporarily stored in an accumulator, so when we need an estimate of how much time has passed – how long you've been reading this article, for example – we access the contents of the accumulator and count up the ticks.

The pacemaker-accumulator model is a "powerful theoretical tool," which is good at explaining how people perform when asked to judge the length of a short interval – the duration of a tone or flashing light, for example. But it runs into serious problems at the physiological level. The idea of a neural 'accumulator' that counts ticks indefinitely is especially problematic. As Buhusi explains, "There's no way for something like this to be coded in the brain."

In recent years, however, neuroscientists have explored the brain's timing mechanisms using measurements of electrical activity and imaging techniques such as positron emission tomography (PET) and magnetic resonance imaging (MRI). They have also studied people whose perception of time has been distorted by brain damage or disease. The result is a more complex model of interval timing known as 'coincidence-detection'.

"This model is much more physiologically realistic," says Warren Meck of Duke University in North Carolina, USA. In late 2005, Meck and Buhusi brought together the theoretical and physiological work in a paper published in Nature Reviews Neuroscience. They proposed a model in which a brain region called the striatum – part of the basal ganglia, deep inside the brain – acts as the control centre for interval timing.

Readers' comments