Thereafter, the pace of record-breaking slowed and, in some cases, stopped completely. For instance, Florence Griffith-Joyner's 10.49 seconds for the 100-metres, set in 1988, remains unchallenged to this day.

In 2007, records were at 99 percent of human physical limits, according to the scientists' statistical model.

The trend was confirmed across a full range of physical exertions, regardless of whether the activity was aerobic (such as the 10,000m speed skating), anaerobic (weightlifting), used lower limbs (cycling) or upper limbs (shot put), explosive (high jump) or endurance (50km walk).

By 2027, athletes will have reached 99.95 per cent of physical limits, meaning that 'half of the world's records will not improve significantly,' the scientists believe.

For example, the men's 100m record of 9.67 seconds will only be improved by tiny slivers of time, measured in the thousands of seconds.

And by 2060, the phenomenon will have spread to all of those disciplines, predicted lead researcher Jean-Francois Toussaint.

'By 2060, assuming that conditions do not change markedly, we will have hit a ceiling, and a new record will typically be an improvement of one two-thousandth over the previous one,' he said.

'It means we can change the unit of measurement to take this into account, for instance using thousandths of a second for the 100m, hundredths of a second for the marathon, or grammes for weightlifting - but then we might have to wait for half a century for the record to be broken.'

Mr Toussaint's study admits that the statistical model may have been distorted by doping. In the 1980s, for instance, Soviet bloc athletes were notorious for use of steroids and other chemicals.

But, the authors argue, this phenomenon proves the point about how close we are to human limits.

Over the past 40 years, the science of training and diet and the social and economic resources devoted to top-level sport have all improved hugely, yet record busting has slowed. Without performance-enhancing drugs, the buffers would have been hit even earlier.

'A fundamental question arises from this,' said Mr Toussaint. 'How do we see the future of sports that depend crucially on record-breaking? How can we, in the light of this, develop a new policy on sports?'

The study appears in the open-access journal PLoS One, published by the US-based Public Library of Science (PLoS). -- AFP