Learning Strategies and Memory Mechanisms

Henrike Hultsch , research associate

A number of facts make the song learning of nightingales a unique paradigm to study song learning. First, males can be tape-tutored, which allows elegant control over what birds hear (e.g. which song patterns, when, how often, in which context). Second, their large vocal repertoire and the acoustic precision of imitations allow to ask questions not only on what they learn (the inventory of behaviour) but also how they use that inventory (the singing program or song sequencing; 1-5). Third, the ‘instruction’ during auditory learning and the ‘retrieval’ of acquired song material during motor learning are temporally clearly segregated, allowing to tease apart the role of memorization and developmental mechanisms in the acquisition process (6). Finally, the motor phase of learning takes about four months in this species, allowing a detailed examination of how representations acquired during perceptual learning translate into singing behaviour (7).

In one of our projects we explore the architecture of sequence representation and its relationships to memory mechanisms during the acquisition of song in young nightingales. After having established basic principles of how heard song sequences are translated into singing behaviours (1-5), we are now focusing on whether sequence acquisition and representation entails cognitive accomplishments, such as the categorization of stimuli according to contextual variables of the tutoring. We examine these questions by labeling particular tutoring contexts (e.g. different social conditions) by particular sets of song stimuli. If categorization occurred during auditory learning, we expect that the birds’ own performances are organized according to the tested contextual variable. Such competence would not only reveal ‘semantic’ properties of the song memory. It may also provide functional explanations for the birds’ vocal interactions in the field.

Currently the lab develops operant conditioning techniques to shift the ‘non-interactive’ tape tutoring to a design where the male has behavioural control over which songs to hear when and how often, or when to switch to novel patterns. Monitoring such decisions and relating them to acquisition success will elucidate the dynamics of the learning process and the underlying strategies. Operant techniques will also be used during vocal ontogeny to study whether and how further auditory experience (additional exposure to song) may influence the quality and progression of the process (8).

(1) Hultsch, H. & D. Todt 1989a. Memorization and reproduction of songs in nightingales (Luscinia megarynchos): Evidence for package formation. J. comp. Physiol. A 165, 197-203.
(2) Hultsch, H. & D. Todt 1989b. Context memorization in the song learning of birds. Naturwissenschaften 76, 584-586.
(3) Hultsch, H. 1992a. Time window and unit capacity: dual constraints on the acquisition of serial information in songbirds. J. Comp. Physiol. A 170, 275-280.
(4) Hultsch, H. & Todt, D. 1992b. The serial order effect in the song acquisition of birds: effects of exposure frequency to models. Anim. Behaviour, 590-593.
(5) Hultsch, H. & Todt, D. 1996 Discontinuous and incremental processes in the song learning of birds: evidence for a primer effect. J. Comp. Physiol. A 179: 191-299.
(6) Hultsch, H. & D. Todt 2004 Learning to Sing. In: Marler, P. & H. Slabbekorn (eds), Nature's Music - The Science of Birdsong, Elsevier Publishers, Academic Press, Amsterdam & Boston, pp. 80-107.
(7) Brumm, H. & Hultsch, H. 2001 Pattern amplitude is relaled to pattern imitation during the song development of nightingales. Animal Behaviour 61: 747-754.
(8) Geberzahn, N. & Hultsch, H. & Todt, D. 2002 Latent song-type memories are accessible through auditory stimulation in a hand-reared songbird. Animal Behaviour 64, 783-790