mantra · om · amygdala

People chant OM out loud. A small scan saw the right amygdala quiet down.

JM Thomas · 21 August 2026

You wanted the syllable to land.

People repeat a mantra when they say one syllable or one short phrase again and again. OM, also written AUM, is the chant a lot of them use. Some say it on a mat, and some say it while they drive. Researchers once put people in a scan that watches blood flow and had them chant it out loud. You came here for that scan.

A drawing of the syllable OM next to the small almond-shaped part of the brain that flags threat.
The syllable OM next to the small part of the brain that flags threat.

What is a mantra?

A mantra is a syllable or phrase you repeat. People who keep a practice often come back to one syllable, and OM, also written AUM, is that syllable for a lot of people. They say the vowel first, then they hold the M, and you can chant it out loud or keep it in the mind. Kalyani's group watched the out-loud version.

What is the amygdala?

The amygdala is a small almond-shaped part of the brain that flags threat. A medical chapter in StatPearls says it is made of about 13 nuclei in five groups, and nuclei here are small clusters of brain cells. You do not need those 13 names to follow this page. You only need the idea that this small part flags threat.

Did a scan watch people chant OM out loud?

Kalyani and colleagues ran that work in 2011 at NIMHANS, the Indian mental-health institute that ran it. They titled the paper Neurohemodynamic correlates of OM chanting, and neurohemodynamic means they watched blood flow in the brain while people chanted. They called the work a pilot, and a pilot is a first small study.

They scanned 12 right-handed healthy people on a 3 Tesla magnet. Tesla is the unit that names how strong the magnet is, and 3 Tesla is a strong hospital magnet. The people were 22 to 39 years old, and the mean age was 28. Nine of the twelve were men, and four already had yoga training.

They used fMRI, a scan that watches blood flow, and each person chanted OM out loud in 15-second blocks. The vowel lasted 5 seconds and the consonant M lasted 10 seconds. They did ten of those blocks, and they put rest blocks between them. They also had people make a 15-second ssss sound, so they could compare a chant to a hiss.

You can read that paper at pmc.ncbi.nlm.nih.gov/articles/PMC3099099.

A drawing of chanting OM, then a quieter small part of the brain that flags threat.
Chanting OM, then a quieter small part of the brain that flags threat.

What did the small OM scan show?

They did not see more blood-flow activity during OM than during rest. They did see deactivation, which means they saw less of that activity than during rest.

The quieter spots were the amygdala, the anterior cingulate, the hippocampus, the insula, the orbitofrontal cortex, the parahippocampal gyrus, and the thalamus. People use those names for parts that help you flag a threat, hold a feeling, and feel the inside of the body. They did not see that change during the ssss sound.

They wrote the right amygdala versus rest at Talairach 24, minus 10, minus 08, and Talairach is a map people use to name a spot in the brain with three numbers. The T score was 5.2, and a T score tells you how far their result was from a no-change line. They used a family-wise error correction, which means they corrected for looking at many spots, and after that correction the p value was less than 0.001. A p value is how often a no-change result would look this strong by luck, and less than 0.001 means that luck is rare. The left amygdala was not in that table. They did not print a percent drop.

Could they just have been resting?

They called it a pilot. They also wrote that a relaxation confound could not be ruled out, and a confound is a second cause that might explain the same picture. They could not prove the quiet came from the syllable alone, and they could not prove it was only rest. You can still want the syllable to land.

Thomas Kraus and colleagues scanned 22 people in 2007 during transcutaneous vagus nerve stimulation, which is a small electrical pulse on the ear, not a chant. They also saw less blood-flow activity in feeling and threat parts of the brain, including the amygdala. Kraus is a neighbor to this page, not this chant.

Did other people watch OM change the body?

Shirley Telles and colleagues watched 7 people in 1995 keep OM in the mind, which is mental OM, not the out-loud version, and they saw a slower heart rate. In 1998 they compared OM to the word One, and both groups had a slower heart and a slower breath. Only the OM group showed lower skin resistance, which is how easily a tiny current crosses the skin.

Luciano Bernardi and colleagues watched 23 people in 2001 say a rosary prayer at six cycles in one minute. A rosary is a repeated prayer on beads. That work is about a repeated phrase and a slow breath, not about OM in a magnet.

What about a sung loop like kirtan?

Kirtan is a sung loop of names or syllables. Open the kirtan page if you want that loop, and stay here if you want the out-loud OM scan.

If chant or this work brings unusual distress, talk to a clinician, because this page is educational.

Can you still want the syllable to land?

You wanted the syllable to land. Keep the mantra name.

People chant OM out loud. You can still want the syllable to land tonight.

Sources

Kalyani B.G. and colleagues, "Neurohemodynamic correlates of OM chanting: A pilot functional magnetic resonance imaging study," International Journal of Yoga 4(1):3-6 (2011), NIMHANS, the Indian mental-health institute that ran it, a pilot, which means a first small study, 12 right-handed healthy people, 3 Tesla, ages 22 to 39, mean 28, nine men, four with yoga training.

People chanted audible OM with the vowel for 5 seconds and the consonant M for 10 seconds, in 15-second blocks, ten blocks, with rest between them, plus a 15-second ssss. No significant activation for OM versus rest. Significant deactivation in the amygdala, anterior cingulate, hippocampus, insula, orbitofrontal cortex, parahippocampal gyrus, and thalamus. They saw neither during ssss. Right amygdala versus rest at Talairach 24, minus 10, minus 08, T of 5.2, family-wise error p less than 0.001. Left amygdala not in that table. Relaxation confound could not be ruled out. No percent drop printed. PMID 21654968, PMC3099099. pmc.ncbi.nlm.nih.gov/articles/PMC3099099

Kraus T. and colleagues, 2007, 22 people, transcutaneous vagus nerve stimulation, which is a small electrical pulse on the ear, not a chant, less blood-flow activity in feeling and threat parts of the brain, including the amygdala. Neighbor to this page, not this chant.

StatPearls, amygdala about 13 nuclei in five groups, and nuclei here are small clusters of brain cells. One sentence on this page.

Telles S. and colleagues, 1995, 7 people, mental OM, which means they kept the syllable in the mind, slower heart rate. Telles 1998, OM versus the word One, both groups had a slower heart and a slower breath, only the OM group showed lower skin resistance, which is how easily a tiny current crosses the skin.

Bernardi L. and colleagues, 2001, 23 people, rosary prayer, which is a repeated prayer on beads, six cycles in one minute. A repeated phrase and a slow breath, not OM in a magnet.