The fourth dimension - time - is essential for many cognitive processes, and for rhythmic movements such as walking. Recent research has begun to elucidate how neuronal activity encodes events that occur on the timescale of tens to hundredths of milliseconds (hundredths of a second) and contain cues which are required for processes such as visual perception, speech discrimination and fine movements.
Many organisms time events on much larger scales. However, next to nothing is known about the mechanisms by which the brain encodes longer periods of time. A new study now sheds some light on these processes. It shows that the rhythmic activity of neuronal ensembles in the zebrafish visual system can encode the time interval of repetitive visual stimuli, and that this metronome-like activity can retain the memory of the interval for up to 20 seconds.
Mu-Ming Poo and his colleagues at the University of California investigated the properties of groups of neurons in a part of the zebrafish visual system called the optic tectum, which is located in the midbrain. The tectum processes visual information and integrates it with inputs from other sensory systems. The equivalent of the tectum in mammals (including humans), a structure called the superior colliculus, is known to be involved in generating saccadic eye movements and likely has other as yet unknown functions.
The researchers loaded optic tectum neurons in zebrafish larvae with a fluorescent calcium-sensitive dye, which can be used to monitor the tiny increases in calcium ion concentratiosn which are associated with neuronal activity. The fish were then presented with visual stimuli consisting of light flashes or bars moving in various directions, and the activity of a population of approximately 200 tectal cells induced by the stimuli was simultaneously observed using confocal and two-photon microscopy.
Visual stimulation of one eye with a moving light bar was found to induce a robust, reliable and synchronous response in some neurons in the tectum on the opposite side of the brain. Other tectal neurons responded only sporadically, or not at all. When the bar stimulus was made to move in the opposite direction, different but partially overlapping ensembles of cells were activated. Synchronized activity was never observed in the absence of the stimulus, suggesting that each ensemble is activated by a stimulus consisting of a specific pattern.
The moving bar stimulus was then presented 20 times in succession, with an interval of 6 seconds between each presentation. Afterwards, a subpopulation of tectal cells exhibited increases in calcium ion concentration at times corresponding to multiples of the interval. These rhythmic calcium currents were observed only in those neurons which had previously responded to the same stimulus, and generally persisted for up to 3 cycles, or 18 seconds, after the final stimulus was removed. The responses were smaller than those evoked by the stimulus itself, and got progressively smaller with each cycle. When the interval between each stimulus was changed to 4 or 10 seconds, the periodicity of the post-stimulus rhythmic activity was found to change correspondingly.
The relevance of this rhythmic neuronal activity to zebrafish behaviour was then explored. Flashes of light were found to evoke tail flipping in larvae whose heads were immobilized in an agarose gel. When presented with 20 flashes, the larvae flipped their tails for up to 20 seconds afterwards, even in the absence of visual stimulation. Significantly, these tail flips occurred at a frequency that corresponded to the interval between the flashes that head been presented just before. Furthermore, towards the end of the time during which the repetitive stimulus was being presented, "anticipatory" tail flips were seen to occur immediately before the onset of the light flashes.
Simultaneous monitoring of the tail flips and calcium currents revealed a strong correlation between the movements and the synchronous activity of ensembles of neurons in the tectum. These findings suggest that the rhythmic activity of tectal neurons contributes to short-term perceptual memories of visual stimuli, which could enable zebrafish larvae to estimate when an impending stimulus is likely to occur. Further work is needed to determine how the neuronal populations in the tectum are connected to the cells which control rhythmic movements.
Sumbre, G. et al (2008). Entrained rhythmic activities of neuronal ensembles as perceptual memory of time interval. Nature DOI: 10.1038/nature07351
Is it possible that these tectal cells are used to record the length between stimuli in short term memory but are not actually co-opted for a general "metronome" in other cases? I mean, what evidence do we have that these tectal cells are part of a general mechanism and not a specific one?
I think the evidence that these cells encode the timing of the visual stimuli used in this study, is strong. But as you say, they might not be involved in a more general metronome mechanism.
Although they're part of thevisual system, tectal cells also encode information from other sensory modalities, so it would be interesting to see how they respond to different types of stimuli (e.g. repetitive sounds).