Science: A Revolution in Neuroimaging! The time resolution is increased by 8 times, and the "artifact" to explore neuronal activity is born!

*For medical professionals only

With the development of functional magnetic resonance imaging (fMRI) technology since the nineties, it has gradually become the main means of non-invasive monitoring of neuronal activity [1].

The imaging principle of fMRI is the blood oxygen level dependent (BOLD) effect, that is, the increase in oxygen consumption during neuronal activity, resulting in an increase in local blood flow, which is an indirect neuronal activity imaging method
.
With high-field strength MRI equipment, fMRI spatial resolution can reach the millimeter level
.

However, the temporal resolution of fMRI is limited by the relatively slow hemodynamic changes caused by neuronal activity, and the detected signal lags about 1 second behind neuronal activity, but the transmission of signals between neurons takes only a few milliseconds [2].

Therefore, by achieving higher temporal resolution (milliseconds), scientists will be able to more precisely reveal the dynamic processes
behind neuronal activity.

Recently, the Jang-Yeon Park team from Sungkyunkwan University in South Korea and the Jeehyun Kwag team from Korea University have published a blockbuster study in the journal Science [3], and they have developed a fast linear scanning method (DIANA) that directly images neuronal activity with millisecond-level accuracy while retaining the high spatial resolution of high-field intensity MRI, providing a new method for in-depth understanding of the spatiotemporal dynamics of neural networks
。

The researchers successfully imaged mouse neuronal activity by the DIANA method at a condition of 9.
4 Tesla (T, magnetic flux density), recording the activity sequence of neurons in the thalamus and primary somatosensory cortex after electrical stimulation of the mouse beard pad with a temporal resolution of 5 milliseconds (8 times that of standard fMRI technology) and a spatial resolution of 0.
22 mm.
It can be called an "artifact"
for studying neural pathways.

Screenshot of the front page of the paper

To achieve high temporal resolution in the millisecond range, the researchers employed a 2D fast linear scan—combined with a linear scan acquisition strategy and a method using a 2D gradient echo imaging sequence with short echo time (TE, 2 ms) and short repeat time (TR, 5 ms) [4], to give 5 Hz electrical stimulation to the left whisker pad of mice placed within a 9.
4T field strength (intensity: 0.
5mA; duration: 0.
5 ms) to image
the plane (layer thickness) of the right primary somatosensory cortex (S1BF) (layer thickness 1 mm).

Schematic diagram of the DIANA method

After electrical stimulation of the beard pad, compared with before stimulation, a significantly enhanced signal can be observed in the right primary somatosensory cortex, and the peak signal appears at 25.
00±1.
58 ms after electrical stimulation, realizing millisecond-level temporal resolution imaging
of neuronal activity.

Millisecond-level temporal resolution imaging of neuronal activity

To confirm the correlation between this signal and neuronal activity, the researchers implanted probes in the primary somatosensory cortex of mice, recorded local potential (LFP) and individual neuronal action potential activity, and administered the same electrical stimulation
to the mice.

The results showed that the peak local potential latency induced by electrical stimulation of the beard pad was 39.
48±1.
84 ms, which was significantly slower than that of the signal in DIANA, but the peak action potential discharge rate latency recorded by the electrode (26.
44±1.
24 ms) was similar
to the signal latency in DIANA.

In addition, the other temporal characteristics of the action potential recorded by the electrode are similar to the signals in DIANA, and the amplitude and rate of the signal in DIANA increase with the increase of electrical stimulation intensity (signal latency is unchanged), which reflects the obvious correlation
between signals and neuronal activity in DIANA.

Signals in DIANA have a clear correlation with neuronal activity

Next, the researchers explored the ability to
capture action potentials propagating between neurons through the DIANA method.

After electrical stimulation of the beard pads, the researchers imaged the coronal level (layer thickness 1 mm) of the thalamus and primary somatosensory cortex, and conventional BOLD-fMRI showed that neurons in both the thalamus and the contralateral/ipsilateral primary somatosensory cortex were activated
.

However, the DIANA method shows the sequence of activation of
brain regions.
After electrical stimulation, the thalamus, contralateral primary somatosensory cortex and ipsilateral primary somatosensory cortex were activated sequentially, and the signal latency was 11.
50±0.
76 ms, 24.
00±1.
00 ms, and 28.
50±3.
08 ms, respectively, that is, thalamic neurons were activated
10 to 15 milliseconds before the contralateral and ipsilateral primary somatosensory cortex.

The DIANA method shows sequential activation of the thalamus, contralateral primary somatosensory cortex, and ipsilateral primary somatosensory cortex after electrical stimulation

To determine whether the spatiotemporal propagation of DIANA signals matches the propagation of neuronal action potentials in thalamic cortical pathways, the researchers probed the simultaneous recording
of electrical activity in the thalamus and primary somatosensory cortex.

The results showed that after electrical stimulation of the beard pad, the activation of neural pathways was indeed consistent with the signals
observed in the DIANA method according to the order of the thalamus (9.
52±0.
90 ms), the contralateral primary somatosensory cortex (24.
52±1.
43 ms) and the ipsilateral primary somatosensory cortex (30.
00±1.
73 ms).

The sequence of neural pathway activation recorded by the probe is consistent with the signal observed in the DIANA method

In addition, the researchers also verified DIANA's ability to
capture neuronal activity through optogenetic methods.

After photostimulation of the primary somatosensory cortex excitatory neurons expressing channel rhodopsin 2 (ChR2) (intensity: 50mW/m㎡; Duration: 20 ms), first signal changes occur in the primary somatosensory cortex (latency 15.
00± 1.
29 ms), followed by signal changes in the thalamus (latency 25.
00 ±2.
98 ms).

This is consistent with neuronal action potentials recorded by the probe after light stimulation (primary somatosensory cortex latency 9.
06±1.
59 ms, thalamus 21.
86±2.
76 ms).

DIANA captures neuronal activity transmission in the corticothalamic pathway after photostimulation

What's even more amazing is that thanks to its high spatiotemporal resolution, DIANA can also distinguish the spatiotemporal order
of activation of substructures in brain regions.
The thalamus can be subdivided into the posteromedial portion (POm), the ventral part of the posterior ventromedial nucleus (VPMv), and the dorsal part of the posterior ventromedial nucleus (VPMd), while the primary somatosensory cortex and S2 (secondary somatosensory cortex) can be divided into L2/3, L4, L5, and L6 layers
.

After electrical stimulation of the beard pad, signals in the thalamus first appear in VPMv and then are transmitted to POm and VPMd; In the primary somatosensory cortex, signals begin at L4 and L5 and then propagate to L6 and L2/3; in the secondary somatosensory cortex, L4 and L6 are activated first, followed by L5 and L2/3
.
These results are consistent
with neuronal action potential results recorded by the probe.

DIANA distinguishes the activation sequence of thalamus, primary somatosensory cortex, and secondary somatosensory cortex substructures

So what is the physiological mechanism of signaling in DIANA?

To be sure, the signal in DIANA does not originate from the BOLD effect
.

To demonstrate this, the researchers performed the BOLD-fMRI experiment under two conditions: an oxygen-to-air ratio of 1:4 (oxygen-air conditions) and air-only conditions
.

Compared with oxygen-air conditions, BOLD effector signals in the thalamus and primary somatosensory cortex were significantly reduced under air-only conditions (thalamus: 0.
626±0.
052% to 0.
284±0.
079%; Contralateral primary somatosensory cortex: 1.
142±0.
147% to 0.
792±0.
166%), consistent
with the dependence of the BOLD effect on oxygen supply.

In contrast, there was no significant change in signal intensity in the DIANA method under oxygen-air and air-only conditions (thalamus: 0.
199±0.
018 to 0.
193±0.
008%; Contralateral primary somatosensory cortex: 0.
164±0.
013% to 0.
165±0.
009%)
.

The signal in the DIANA method does not originate from the BOLD effect

While the researchers don't fully understand the physiological mechanisms behind this signal, the researchers suspect that the signal may be related to
changes in membrane potential during neuronal activity.

In general, the researchers developed a method to improve the temporal resolution to the millisecond level, while retaining the high spatial resolution of non-invasive neuronal activity imaging method, to achieve high temporal and spatial resolution detection of brain substructure of mouse brain neuronal activity, which provides an important method
for future neuroscience research.

ADtaxi-Memories Vol.
05 - Professor Henrik Zettberg, Institute of Neuroscience and Physiology, University of Gothenburg, Sweden, takes you to understand the future pattern of AD biomarkers

However, there is still a question mark
over whether this method is feasible in humans.
You know, it is difficult for people to remain inactive for a long time in a waking state, and how to reduce the artifacts caused by exercise is a technical problem
that the DIANA method cannot bypass.

However, everything is difficult at the beginning, and with the advancement of technology, this method will eventually play an important role
in future scientific research and clinical practice.

References

1.
Belliveau JW, Kennedy DN, Jr.
, McKinstry RC, Buchbinder BR, Weisskoff RM, Cohen MS, Vevea JM, Brady TJ, Rosen BR: Functional mapping of the human visual cortex by magnetic resonance imaging.
Science 1991, 254(5032):716-719.

2.
style="white-space: normal;margin: 0px;padding: 0px;box-sizing: border-box;text-align: left;">3.
Toi PT, Jang HJ, Min K, Kim S-P, Lee S-K, Lee J, Kwag J, Park J-Y: In vivo direct imaging of neuronal activity at high temporospatial resolution.
Science 2022, 378(6616):160-168.

4.
Yu X, He Y, Wang M, Merkle H, Dodd SJ, Silva AC, Koretsky AP: Sensory and optogenetically driven single-vessel fMRI.
Nat Methods 2016, 13(4):337-340.

Responsible EditorBioTalker