Neurophysiology of Vocal Learning, Multimodal Processing & Neuromodulation Lab

The Speech and Language Center of the Department of Neurology conducts basic neurophysiological studies in the songbird zebra finch, and functional magnetic resonance imaging (fMRI) and brain magnetic stimulation studies in humans. The songbird research pertains to the neurobiology of vocal learning and animal models of speech-motor disorders. Under projects supported by NIDCD, NIMH and NSF we have made several key observations: During the normal development of birdsong in zebra finches, variant birdsongs exist that contain repetitive vocal patterns resembling dysfluencies of stuttering (figure1); modifying the social environment (e.g., the birds with phonatory iterations are exposed to birds singing normal bird song) produces some recovery toward normal singing patterns.

Figure 1. Normal and variant birdsongs
A. Spectrogram of a normal zebra finch song.
B. Spectrogram of a variant song containing syllable repetitions (R).
C. A male-female zebra finch pair (Helekar, Salgado-Commissariat Rosenfield and Voss (2013).

Our behavioral and brain slice experiments indicate bidirectional modulation of specific brain receptors (neuronal nicotinic cholinergic receptors) play a critical role in vocal learning and variant song production. In collaboration with Dr. Henning Voss (Weill Medical College of Cornell University), we have adapted functional magnetic neuroimaging technology to zebra finches. These studies reveal a possible deficiency in sensory learning of tutor song in the songbird model of dysfluency (Figure 2), and stimulus- and context-specific experience-dependent plasticity of the fMRI response in auditory brain areas. In collaboration with Dr. Ofer Tchernichovski (Hunter College, New York) we developed a Balloon-Bird wireless recording procedure to record local field potentials simultaneously from several forebrain structures during audiovisual video stimulation in freely moving zebra finches (Figure 3).

Our human fMRI studies focus upon brain multimodal, biological motion and mirror neuron processing systems in normal subjects and people who stutter. We are using traditional general linear model-based methods of analysis of fMRI data, and also contributing to the development of functional connectivity and functional spatial mapping methods.

Figure 2. fMRI in birds with variant songs (repeaters) and normal songs (non-repeaters)
Mean BOLD activation for four different auditory stimuli - tutor song (TUT), 2 kHz pure tone (TONE), bird’s own song (BOS) and conspecific song (CON) for 8 non-repeater controls birds (Non-repeaters), and 8 repeater birds (Repeaters). Colors denote correlation coefficients R, individually scaled in each plot, overlaid to averaged EPI images (grey). White border depicts the brain template outline. The main activated area that is consistently activated in all images corresponds to NCM, CM and field L region. Difference images represent the difference in z values between non-repeaters and repeaters. (Helekar, Salgado-Commissariat and Voss, 2010).

Our recent fMRI studies on stuttering subjects viewing video clips of speech reveal altered activation of parietal multimodal and mirror neuronal areas, and changes in resting state functional connectivity of speech-specific areas. Other studies comparing trained pianists with matched untrained control subjects show selective bilateral activation of visual motion-sensitive areas V5/MT and pSTS in subjects viewing video clips of finger movements on piano keys rendering a musical melody. We find that the activation is enhanced in trained pianists compared to controls, and that the measured increase scales with increasing duration of experience as a pianist for up to 22 years (Figure 4). These results suggest visual motion sensitive areas can undergo experience-dependent functional neuroplasticity. We determined that in both trained pianists and controls watching and hearing the performance of a musical melody on a piano correlates with bilateral activation of occipital-temporal and parieto-temporal multimodal areas. However, the magnitude and volume of activation of these brain areas is substantially greater in piano-trained subjects compared to untrained subjects, suggesting again these areas are subject to functional neuroplastic changes.

Figure 3. Recording of LFP responses to audiovisual call stimuli
A. Picture showing an awake freely moving bird during recording with the wireless headstage and helium balloon attached to its head.
B. A diagram of the zebra finch brain showing labeled structures in which recording electrodes were placed.
C. Traces showing LFPs recorded from the labeled structures (NCM, L, W and E refer to caudal medial nidopallium, field L, visual wulst and entopallium, respectively). V and A on the left indicate durations of the visual and auditory components of the audiovisual stimulus, respectively. AV, A and V on the top point to responses to audiovisual, audio and visual stimuli, respectively.

Regarding brain stimulation studies, Drs. Helekar and Voss developed a wearable multifocal transcranial rotating permanent magnet stimulator (TRPMS) that can be triggered wirelessly in potential in-home non-invasive neuromodulatory therapy, or whose spatiotemporal stimulus parameters can be controlled interactively and dynamically, whenever necessary, in human neuroscience research.

Figure 4. Plasticity of the visual motion-sensitive areas
Significant BOLD responses (p ≤ 0.05 Family-wise error-corrected t test) associated with 40 second video with no audio stimulation of moving fingers on the piano (V) minus (i.e. greater than) a still frame of fingers on the piano of the same duration (S) contrast (V – S) in trained pianists (n = 17) and untrained control subjects (n = 16). Glass brain and 3D surface representations of activated areas are shown.