Technical Comments

Comment on "Cerebellar LTD and Learning-Dependent Timing of Conditioned Eyelid Responses"

See allHide authors and affiliations

Science  09 Apr 2004:
Vol. 304, Issue 5668, pp. 211
DOI: 10.1126/science.1093706

In a recent report on eyeblink conditioning in mice, Koekkoek et al. (1) drew several conclusions that we believe are unsubstantiated and incorrect and fail to cite much of the relevant literature in this field.

First, Koekkoek et al. argued that their magnetic method of measuring eyelid movement is superior to electromyogram (EMG) recording from the obicularis oculi. We previously showed that eyelid movement is essentially perfectly correlated with EMG measurement for conditioned eyeblink responses (CRs) (2, 3). EMG recording is slightly more sensitive, particularly for measuring CR onset latency, and is the current standard in freely moving animals. The only advantage of the method used in (1) is more accurate measurement of eyelid response to unconditioned stimulus (US), which is irrelevant to the study.

Second, on the basis of poorly placed lesions of the interpositus (IP) nucleus, Koekkoek et al. claimed that IP lesions do not abolish the eyeblink CR. They further concluded that the cerebellum is not necessary for eyeblink conditioning. However, their data are consistent with the residual CRs observed following partial lesions (4). Koekkoek et al. (1) also did not mention post-lesion CR percentages, which is troubling in light of low rates of pre-lesion CRs in mutants. Perhaps most puzzling is the remarkable decrease in CR onset and peak latencies in wild-type mice following IP lesions. Although not explained in (1), it is striking that the most significant change in CR timing occurs in wild-type mice with IP lesions—an effect that cannot be attributed to compromised cortical long-term depression (LTD).Koekkoek et al. failed to cite numerous studies from our laboratory and others showing that appropriately placed IP lesions completely prevent and abolish the conditioned eyeblink response, as well as extensive literature supporting the view that the cerebellum and its associated circuitry is the essential substrate for delayed classical conditioning of eyeblink and other discrete responses (5). In particular, appropriate bilateral IP lesions completely prevent learning and completely abolish the learned response in the freely moving mouse (6-8).

Third, using a mutant mouse with impaired LTD, Koekkoek et al. reported that these animals have shorter than normal latency CRs and argue that LTD is critical for CR timing (9). Earlier studies have shown that appropriate lesions of the cerebellar cortex result in shorter latency CRs (5, 10, 11). In contrast, in collaborative studies we showed that LTD is closely correlated with normal learning: Impaired LTD is associated with impaired eyeblink conditioning (12-14). Importantly, mutant mice with no cerebellar LTD, although impaired in learning, show normal eyeblink CR onset and peak latencies (13). The correlation claimed by Koekkoek et al. thus appears to be spurious. Nonetheless, LTD and other processes of cortical plasticity may all play roles in CR timing, given the cortical lesion effect.

In addition to the problems noted above, Koekkoek et al. did not state their criteria for defining CRs. Many of the CRs reported in (1) have much too short onset latencies to be CRs—and may actually have been startle responses—which would invalidate conclusions regarding CR timing. Furthermore, the CR percentage of both wild type and L7-PKCi mutant mice on training day 2 (T-2) is extremely high and at odds with all the published literature, as is the reported ∼50% decrease in CR rate at T-3 and T-4 [figure 1 in (1)]. Such decreases resemble extinction-like response rates and suggest that the mutants may have unidentified deficits in sensitivity to the US or altered climbing fiber— Purkinje cell plasticity. No controls were run to assess possible alterations in sensory sensitivity (to tone CS and shock US). These controls are necessary when using mutant mice (6, 15).

References and Notes

Navigate This Article