Ryusuke Kakigi

In music, melodic information is thought to be encoded in two forms, a contour code (up/down pattern of pitch changes) and an interval code (pitch distances between successive notes). A recent study recording the mismatch negativity (MMN) evoked by pitch contour and interval deviations in simple melodies demonstrated that people with no formal music education process both contour and interval information in the auditory cortex automatically. However, it is still unclear whether musical experience enhances both strategies of melodic encoding. We designed stimuli to examine contour and interval information separately. In the contour condition there were eight different standard melodies (presented on 80% of trials), each consisting of five notes all ascending in pitch, and the corresponding deviant melodies (20%) were altered to descending on their final note. The interval condition used one five-note standard melody transposed to eight keys from trial to trial, and on deviant trials the last note was raised by one whole tone without changing the pitch contour. There was also a control condition, in which a standard tone (990.7 Hz) and a deviant tone (1111.0 Hz) were presented. The magnetic counterpart of the MMN (MMNm) from musicians and nonmusicians was obtained as the difference between the dipole moment in response to the standard and deviant trials recorded by magnetoencephalography. Significantly larger MMNm was present in musicians in both contour and interval conditions than in nonmusicians, whereas MMNm in the control condition was similar for both groups. The interval MMNm was larger than the contour MMNm in musicians. No hemispheric difference was found in either group. The results suggest that musical training enhances the ability to automatically register abstract changes in the relative pitch structure of melodies.

We recorded evoked potentials (EPs) induced by conventional transcutaneous electrical stimulation (TS), laser stimulation (LS) and epidermal electrical stimulation (ES) using a specially made needle electrode. We evaluated the activated fibers by epidermal stimulation by assessing the conduction velocity (CV) of the peripheral nerves. The EPs were recorded from Cz electrode (vertex) of the International 10-20 system in 12 healthy subjects. For the ES, the tip of a stainless steel needle electrode was inserted in the epidermis of the skin (0.2 mm in depth). Distal and proximal sites of the upper limb were stimulated by the LS and ES with an intensity which induced a definite pain sensation. Similar sites were stimulated by TS with an intensity of two times the sensory threshold. A major EP positive response (P1) was obtained by stimulation by all three types of stimuli. The P1 latency for the TS (245+/-22 ms) was significantly shorter than that for the ES (302+/-17 ms, P<0.0001) and LS (341+/-21 ms, P<0.0001) and the peak latency P1 by the LS was also significantly longer, approximately 40 ms, than that by the ES (P<0.0001). The CVs were 15.1, 15.3 and 44.1 m/s obtained by ES, LS and TS, respectively. The CV indicated that the fibers activated by the ES were mainly A fibers, which corresponded to the fibers stimulated by the LS. We considered that the ES with our newly developed needle electrode was a very convenient method for the selective stimulation of the A fibers, since it was very simple, not requiring any special apparatus, did not cause bleeding or burns and caused minimum uncomfortable feeling.

Whether there is an absolute critical period for acquiring language is a matter of continuous debate. One approach to address this issue is to compare the processes of second language (L2) learning after childhood and those of first language (L1) learning during childhood. To study the cortical process of postchildhood L2 learning, we compared event-related brain potentials recorded from two groups of adult Japanese speakers who attained either high or intermediate proficiency in English after childhood (J-High and J-Low), and adult native English speakers (ENG). Semantic anomalies embedded in English sentences evoked a clear N400 component in all three groups, with only the time course of the brain activation varying among the groups. Syntactic violations elicited a left-lateralized negativity similar to the left anterior negativity in ENG and J-High, but not in J-Low. In ENG, a P600 component was additionally found. These results suggest that semantic processing is robust from early on in L2 learning, whereas the development of syntactic processing is more dependent on proficiency as evidenced by the lack of the left-lateralized negativity in J-Low. Because early maturation and stability of semantic processing as opposed to syntactic processing are also a feature of L1 processing, postchildhood L2 learning may be governed by the same brain properties as those which govern childhood L1 learning. We argue that these processes are qualitatively similar in many respects, with only restricted domains of language processing being subject to absolute critical period effects.