A Plausible Mechanism for Drosophila Larva Intermittent Behavior
289
been extended to encompass distinct behavioral modes bearing different go/turn
parameters, thus termed composite Levy walk. Levy walks have been extensively
studied in the context of optimal foraging theory. A Levy walk with a power-law
exponent between the limit of ballistic (α = 1) and brownian motion (α = 3)
yields higher search efficiency for foragers with an optimum around α = 2 when
search targets are patchily or scarcely distributed and detection of a target halts
displacement (truncated Levy walk) [5].
Nevertheless, the underlying assumption of non-intermittent flow of movement in Levy walk models complicates the identification of the underlying
generative mechanisms as they focus predominantly on reproducing the observed
spatial trajectories, neglecting the temporal dynamics of locomotory behavior.
Therefore, Bartumeus (2009) stressing the need for a further extension coined
the term intermittent random walk, emphasizing the integration of behavioral
intermittency in the theoretical study of animal movement [2]. Here we aim
to contribute to this goal by studying the temporal patterns of intermittency
during Drosophila larva free exploration in experimental data and in a conceptual model, bearing in mind that power-law like phenomena can arise from a
wide range of mechanisms, possibly involving processes of different timescales
[5]. While our study remains agnostic towards whether foragers really perform
Levy walks - a claim still disputed [5] - we suggest that intrinsic motion intermittency should be taken into account and the assumption of no pauses and acute
reorientations should be dropped in favor of integrative models encompassing
both activity and inactivity.
Drosophila larva is a suitable organism for the study of animal exploration
patterns and the underlying neural mechanisms. A rich repertoire of available
genetic tools allows acute activation, inhibition or even induced death of specific
neural components. Crawling in 2D facilitates tracking of unconstrained behavior. Also, fruit flies during this life stage are nearly exclusively concerned with
foraging. Therefore a food/odor-deprived environment can be largely considered
stimulus-free, devoid of reorientation or pause sensory triggers, while targetdetection on contact can be considered certain. Truncated spatial Levy-walk
patterns of exploration with exponents ranging from 1.5 to near-optimal 1.96
that hold over at least two orders of magnitude have been previously reported
for the Drosophila larva. The turning-angle distribution, however, was skewed
in favor of small angles and a quasi-uniform distribution was observed only for
reorientation events ≥50
◦ [9]. Moreover, it has been shown that these patterns
arise from low-level neural circuitry even in the absence of sensory input or
brain-lobe function and have therefore been termed ‘null movement patterns’
[8,9].
Behavioral intermittency has not been described for the fruitfly larva. Previous empirical studies on adult Drosophila intermittent locomotory behavior have
concluded that the distribution of durations of rest bouts is power-law while
that of activity bouts has been reported to be exponential [12] or power-law
[8]. Genetic intervention has revealed that dopamine neuron activation affects
the activity/rest ratio via modulation of the power-law exponent of the rest
289
been extended to encompass distinct behavioral modes bearing different go/turn
parameters, thus termed composite Levy walk. Levy walks have been extensively
studied in the context of optimal foraging theory. A Levy walk with a power-law
exponent between the limit of ballistic (α = 1) and brownian motion (α = 3)
yields higher search efficiency for foragers with an optimum around α = 2 when
search targets are patchily or scarcely distributed and detection of a target halts
displacement (truncated Levy walk) [5].
Nevertheless, the underlying assumption of non-intermittent flow of movement in Levy walk models complicates the identification of the underlying
generative mechanisms as they focus predominantly on reproducing the observed
spatial trajectories, neglecting the temporal dynamics of locomotory behavior.
Therefore, Bartumeus (2009) stressing the need for a further extension coined
the term intermittent random walk, emphasizing the integration of behavioral
intermittency in the theoretical study of animal movement [2]. Here we aim
to contribute to this goal by studying the temporal patterns of intermittency
during Drosophila larva free exploration in experimental data and in a conceptual model, bearing in mind that power-law like phenomena can arise from a
wide range of mechanisms, possibly involving processes of different timescales
[5]. While our study remains agnostic towards whether foragers really perform
Levy walks - a claim still disputed [5] - we suggest that intrinsic motion intermittency should be taken into account and the assumption of no pauses and acute
reorientations should be dropped in favor of integrative models encompassing
both activity and inactivity.
Drosophila larva is a suitable organism for the study of animal exploration
patterns and the underlying neural mechanisms. A rich repertoire of available
genetic tools allows acute activation, inhibition or even induced death of specific
neural components. Crawling in 2D facilitates tracking of unconstrained behavior. Also, fruit flies during this life stage are nearly exclusively concerned with
foraging. Therefore a food/odor-deprived environment can be largely considered
stimulus-free, devoid of reorientation or pause sensory triggers, while targetdetection on contact can be considered certain. Truncated spatial Levy-walk
patterns of exploration with exponents ranging from 1.5 to near-optimal 1.96
that hold over at least two orders of magnitude have been previously reported
for the Drosophila larva. The turning-angle distribution, however, was skewed
in favor of small angles and a quasi-uniform distribution was observed only for
reorientation events ≥50
◦ [9]. Moreover, it has been shown that these patterns
arise from low-level neural circuitry even in the absence of sensory input or
brain-lobe function and have therefore been termed ‘null movement patterns’
[8,9].
Behavioral intermittency has not been described for the fruitfly larva. Previous empirical studies on adult Drosophila intermittent locomotory behavior have
concluded that the distribution of durations of rest bouts is power-law while
that of activity bouts has been reported to be exponential [12] or power-law
[8]. Genetic intervention has revealed that dopamine neuron activation affects
the activity/rest ratio via modulation of the power-law exponent of the rest