It's almost impossible to creep up on some animals, such as goldfish, that streak for safety when alarmed. Donald Faber from the Albert Einstein College of Medicine, USA, explains that the fish curl into a tight C shape and zip off in the opposite direction from a threatening sound. The big question was how do fish tell which direction the sound is approaching from to orchestrate the response. According to Faber, fish cannot use the time difference between a sound arriving at both ears to identify the direction, because they are transparent to sound waves and the sound arrives at both ears at the same time. However, he had a hunch that the fish's lateral line – a line of vibration sensors running along the fish's side – may help them to determine the origin of a threatening sound.

Mana Mirjany, Thomas Preuss and Faber designed their experiments to test this idea by taking advantage of the fish's natural behaviour (p. 3358). After inactivating the lateral line of a goldfish with cobalt chloride, Mirjany released the fish to swim naturally in a large circular tank equipped with two loudspeakers. Playing alarming sounds from each of the loudspeakers at random intervals, the team filmed the startled fish's reactions. Repeating the experiment using other techniques to inactivate the lateral line and also blindfolding the fish with custom-made eye caps, Mirjany then painstakingly analysed the fish's escape behaviour when it happened to be in open water and away from the tank sides.

Not surprisingly, the fish turned and fled, regardless of which sense they were deprived of: ‘The auditory system is sufficient to trigger an escape,’ says Mirjany.

Yet, the lateral line was essential for the fish to figure out which direction the threat was coming from, as the fish that had lost their lateral lines fled in random directions whereas the blindfolded fish (with an intact lateral line) successfully headed in the opposite direction from the threat. Yet, when Mirjany detached the body portion of the fish's lateral line from the nervous system, the fish were still able to escape in the opposite direction. So it was the anterior (head) section of the lateral line that was essential for the fish to locate the source of the scary sound and escape correctly.

However, Faber explains that the fish's startle response becomes more complex when the animals are close to an object that could block their escape. Instead, they override the startle response and turn towards the threat in order to avoid colliding with the obstacle. The team decided to find out how the lateral line affected the fish's escape response when near an obstruction.

Analysing the fish's reactions when they happened to be near the tank wall, the team realised that the lateral line and visual systems were both playing crucial roles in determining the direction that the escaping fish chose. The animals successfully overrode their standard escape response when their lateral line was inactivated and, when Mirjany repeated the experiment with blindfolded fish, they too successfully overrode the escape response. It was only when Mirjany blindfolded fish with inactivated lateral lines that the fish lost the ability to override the reflex. ‘There is either compensation between the visual and lateral line systems or there could be integration of the two, but it is difficult to say which at the moment,’ says Mirjany.

Having confirmed that goldfish use the lateral line to pin down the direction of a sound, Mirjany and Faber decided to determine how signals from the anterior portion of the lateral line contribute to the fish's escape response. Explaining that the tightly choreographed escape sequence is coordinated by a single neuron – the Mauthner cell (M-cell), which integrates all of the fish's sensory inputs and determines which direction the fish should flee – Faber and Mirjany decided to analyse the nerve input from the lateral line to the M-cell (p. 3368).

Stimulating the lateral line nerves electrically and recording the M-cell responses at various sites along the neuron, Mirjany found that the M-cell responded within 1 ms. ‘There is only about 3–4 ms from the time the stimulus occurs to the time that one of the Mauthner cells fires an action potential to trigger the response,’ explains Faber, so the M-cell response to the lateral line input was fast enough to control the fish's swift reaction. Also, the lateral line input was close to the M-cell cell body, allowing the coordinating cell to integrate the lateral line inputs with inputs from other sensory systems further out along the M-cell dendrites, to determine which escape strategy is best for each individual situation.

• © 2011.