Three-dimensional holographic photostimulation of the

dendritic arbor
Sunggu Yang, Eirini Papagiakoumou, Marc Guillon, Vincent de Sars,
Cha-Min Tang, Valentina Emiliani

To cite this version:

Sunggu Yang, Eirini Papagiakoumou, Marc Guillon, Vincent de Sars, Cha-Min Tang, et al.. Three-
dimensional holographic photostimulation of the dendritic arbor. Journal of Neural Engineering, IOP
Publishing, 2011, 8 (4), pp.046002. �hal-01976677�

HAL Id: hal-01976677

https://hal.archives-ouvertes.fr/hal-01976677
Submitted on 10 Jan 2019

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 3D Holographic Photostimulation of the Dendritic Arbor

Sunggu Yang 1, Eirini Papagiakoumou2, Marc Guillon2, Vincent de Sars2,

Cha-Min Tang1, and Valentina Emiliani2§

1
Department of Neurology, University of Maryland School of Medicine, 655 West Baltimore St,
and Baltimore VAMC, 10 N. Greene St, Baltimore, Maryland 21201 USA

2
Neurophysiology and New Microscopies Laboratory, Wave front engineering microscopy
group, CNRS UMR 8154, INSERM U603, Paris Descartes University, 45 Rue des Saints Pères,
75270 Paris Cedex 06, France

§ Corresponding author: [email protected]

Abstract:
Digital holography is an emerging technology that can generate complex light patterns for
controlling the excitability of neurons and neural circuits. Strengths of this technique include the
high efficiency with which available light can be effectively utilized and the ability to deliver
highly focused light to multiple locations simultaneously. Here we demonstrate another power of
digital holography, the ability to generate instantaneous three dimensional light patterns. This
capability is demonstrated with photolysis of caged glutamate on the dendritic arbor of
hippocampal neurons to study the nature of integration of inputs arriving on multiple dendritic
branches.
1. Introduction:
Light can be a powerful medium with which the excitability of the nervous system can be
probed [1]. When used in conjunction with caged compounds or in optogenetically engineered
tissue, light can provide the means to control the strength, timing, and location of stimulation
with unparalleled precision. Such control can be achieved without direct contact of the tissue and
with little signal crosstalk.

1
 An ideal light-delivering method should be one that is efficient (minimal power loss
throughout the system), flexible (capable of targeting a wide range of samples, from a single
dendritic spine to a whole 3D population of neurons), and has high resolution in time and space
(millisecond and micrometric, respectively). Several solutions have been recently proposed to
fulfill these requirements. They can be divided in two main categories: laser scanning and
parallel excitation methods. Each of these approaches has advantages and limitations.
Laser scanning methods use galvanometric mirrors or acousto-optic devices (AOD) to quickly
scan a laser beam across several positions [2-5]. One limitation of the laser scanning approach is
the time it takes to scan through a sequence of locations when the geometry is complex and when
it involves a large number of spots.
This can be especially problematic for photolysis of caged compounds because a finite dwell
time is required at each spot regardless of the available power of the light source. The dwell time
is a consequence of inefficient photolysis of available caged compounds [4], particularly in two-
photon [6, 7], depletion of available caged compounds from a very small volume, and light
intensity limitation set by the photodamage threshold of the tissue [8]. The dwell time problem is
accentuated when there is a need to stimulate a large number of synapses. Under a wide range of
physiological conditions large numbers of synapses on a single neuron must be activated within a
brief temporal window to generate an action potential. If the inputs are asynchronous they are
then typically summed linearly or sub-linearly even when spatially clustered, while supra-linear
summation occurs only in the case of near simultaneous arrival of stimuli [9]. Temporal
coincidence of paired excitation is also critical for the induction of synaptic and dendritic
plasticity [10]. Therefore, a method that can provide excitation at a large numbers of sites
simultaneously will be helpful for studying plasticity as well as dendritic integration.
With parallel methods all the regions in the target area are excited simultaneously. They are
based either on the modulation of light intensity (via digital micromirror devices (DMD), liquid
crystal (LC) and light-emitting diode (LED) arrays [11]) or on the modulation of the phase of a
coherent beam (via nematic or ferroelectric liquid crystal arrays). DMD is an example of
amplitude modulating device [12-16]. A DMD is a semi-conductor chip with hundreds of
thousands of switchable micromirrors that can be digitally controlled to produce complex two
dimensional patterns. Its strength includes tolerance to high light intensity and compatibility with
a wide range of wavelength, as well as switching times in the tens of microseconds range.
Nematic- or ferroelectric-LC allows controlling light distribution at the focal plane by a pure

2
phase modulation following the notion of image reconstruction in holography. Originally
proposed for generating multiple traps in optical tweezers [17-21], the experimental scheme for
holographic light patterning consists in computing with an iterative algorithm the phase pattern
at the rear aperture of the objective that reproduces the desired target intensity in the objective
focal plane.
With amplitude modulation, light patterns are created by selectively blocking light, while with
phase modulation light patterns are creating by “redirecting” the light onto the target [22]. The
result is a marked difference in light utilization efficiency when generating sparsely distributed
light patterns.
Another significant difference between these two technologies is the ability to generate 3D
patterns. In biology the structures of interest are normally three dimensional. But current imaging
and optical stimulation approaches are fundamentally two dimensional technologies. While they
can be modified to handle 3D structures, this is generally achieved at a cost of speed and
efficiency [23]. Biologists have responded to this application-technology mismatch by either
forcing the tissue into a 2D state (i.e. culturing it on a glass coverslip) or placing constraints on
the parameter that is being measured. For example, the dendritic arbor of a neuron is in general a
complex 3D structure so that dendritic integration with 2D approaches requires confining the
excitation to relatively short portion of a dendrite which lies in the imaging plane of the
microscope. This constraint makes it difficult to address important questions such as the nature
of multi-branch integration and integration over an extended length of a single dendrite.
On the contrary, holographic light patterning is inherently a 3D technology. Holographic phase
modulation through liquid crystal spatial light modulators (LC-SLM) was originally proposed for
generation of 3D optical traps [17-21]. Its use for 3D imaging or photoactivation has been
recently proposed and discussed [24-26] and the generation of 3D focused spots has been
demonstrated by exciting fluorescent test samples [26, 27]. However, a demonstration for 3D
photoactivation in a biological preparation has not yet been reported. Here, we present and
characterize a holographic microscope system for 3D photolysis and demonstrate applications in
experimental neuroscience with photolysis of caged glutamate on the complex 3D architecture of
the dendritic neurons.

2. 3D holographic spot generation

2.1 Experimental set up

3
Three dimensional control of multiple spot generation is achieved by using a conventional
holographic configuration [24, 28]. Briefly, the output beam of a 150 mW, 405 nm diode laser
(CNI Laser) is expanded by a beam expander (3X) to fill the short axis of a reflective Spatial
Light Modulator (SLM) (LCOS Hamamatsu, model X10468-05). The SLM plane is imaged at
the back aperture of the microscope objective through a telescope (L1, f1=750mm; L2,
f2=500mm). The magnification of the telescope is chosen in order to match the SLM short axis
with the diameter of the objective’s back aperture (Olympus, 60x, W 0.9NA). The undiffracted
component (zero order spot) is removed by placing an aluminum block at the focal plane of L1.
For the photolysis experiments, the holographic illumination system was combined with
electrophysiological recording and conventional fluorescence imaging. Temporal control of the
light pulse was accomplished by gating the laser.
The holographic beam was brought into the optical axis of an upright fluorescence microscope
(Olympus BX51) below the epi-fluorescence unit, with a longpass dichroic mirror. The
intracellular solution of the whole cell patch clamp electrode contained Alexa594 (50µM) or
Cascade Blue hydrazide (750µM)) to identify the location of the thin terminal dendrites and
dendritic spines. Internal solution consisted of (in mM): 135.0 K-gluconate, 5 KCl, 1 MgCl2,
0.02 CaCl2, 0.2 EGTA, 10 HEPES, 4 Na2-ATP, and 0.3 Na-GTP. External solution consisted of
(in mM): 124 NaCl, 4 KCl, 1.23 NaH2PO4, 1.5 MgCl2, 2.5 CaCl2, 26 NaHCO3 and 10 glucose.
Caged glutamate (MNI-Glu, 1-2 mM) was bath applied and recirculated at a rate of 0.7 ml/min.
Experiments were done on hippocampal acute slices, at room temperature. Data were acquired
with pClamp10 software.

2.2 Phase hologram calculation

To locate the excitation spot on the sample plane and calculate the phase holograms, we used a
custom designed software written in C++ with the open graphic library Qt 4.4.0 [24].
In order to minimize the effects of spherical aberrations on the spots generated out of the focal
plane, for the calculation of phase profiles we included in the description of the microscope
objective the non parabolic terms. Briefly, for a high numerical aperture lens satisfying quasi-
stigmatism sine condition, a ray coming from a source placed at the focus and forming an angle
a with the optical axis, will emerge from the lens back aperture at distance d=f·sin(a) with a

phase Dj = 2p .m f , m is the index of refraction of the embedding medium and f is the focal
λ

4
length of the lens in the medium, and l the optical wavelength. Any ray coming from the focus
will leave the lens with the same phase, yielding a plane wave. The lens thus, flattens and
projects the incoming spherical wave of radius f on the back aperture. If we, now, consider
(figure 1 (a)) a ray coming from a point source N (coordinates xn, yn, and zn) placed out of focus
in the sample volume, and reaching the “focal sphere” [29] at a point P (coordinates xp, yp, zp with
2 2
zp = f 2 - x p - y p ), we can write its phase at the focal sphere, and thus at the lens back

2p .m 2p .m
aperture, as Dj pn = NP = ( x p - xn ) 2 + ( y p - y n ) 2 + ( z p - z n ) 2 .
λ λ
Inversely, back propagation of a ray having the former phase at the back aperture of the
objective will give a spot at point N. Let’s then define u p = u exp(if p ) the complex electric field
at the pth pixel of the SLM, imaged in the back aperture of the objective, and use the scalar
diffraction theory to propagate this field distribution, through an objective of focal f, to the
location of the nth spot at the sample space. The complex amplitude vn at the nth spot can be
found as the sum of the N contributions from all the pixels of the SLM:
d2 ( )
å ue
i f p -Df pn
vn = , for small defocus compared to the objective focal length ( z n << f ) the
iλf p =1, N

expression of Djnp can be simplified to: Dj pn =

2p .m
λf
(
1 + xn x p + yn y p + zn f 2 - x 2p - y 2p ) (1),

which for low numerical aperture objectives, , ( x 2p + y 2p << f ) can be further simplified in:

2p .m 2p .m
Dj pn = + (xn x p + yn y p ) - p .m2 zn x 2p + y 2p (2); the last expression, currently used in the
( )
λf λf λf
prisms and lenses algorithm [30], permits the generation of good optical spots, but generates
spherical aberrations for excitation spots out of the objective focal plane with consequent
deterioration in intensity and axial resolution. It was shown that the approximation (2) induced
significant point spread function degradation above 10 µm while modeling high numerical
aperture lenses by equation (1) allowed defocusing of 40 µm or larger [31]. In order to optimize
these parameters, in the calculation of vn we therefore used the expression (1), which hardly
increased computation time.
Finally, to optimize the uniformity of light intensity in multiple spots configuration, we used
the weighted Gerchberg-Saxton algorithm (GSW) [32]. This multi-iterative algorithm consists in

5
attributing a correction weight to the target spots in order to balance numerically-generated
inhomogeneities.
2.3 Optical characterization
In order to characterize the quality of holographic spots organized in 3D structures, we
implemented the holographic scheme, described in section 2.1, on a double microscope where
the upper objective (60x, W 0.9 NA) is used for holographic excitation and the lower one for
imaging (60x, W 1.2NA) [24]. The imaging objective is fixed and focused on a thin fluorescent
layer while the upper one is moved along the axial direction with a piezo scanning stage working
in closed-loop (MIPOS100SG, Piezosystem Jena).
An example of a 3D holographic pattern of diffraction-limited spots arranged in a cubic
structure is shown in figure 1(b). The pattern is shown from two perspectives (side and top
views) to better illustrate the light distribution at the different spot along the axial and transversal
plane, respectively.
In order to define the excitation volume available for photolysis experiments, we characterized
the full width at half maximum (FWHM) and intensity distribution of the holographic spots as a
function of their lateral, dx, and axial, dz, position. To characterize these parameters for the laser
excitation wavelength, 405nm, we replaced the fluorescent layer with a transparent coverslip and
directly imaged the holographic spots through the lower objective into the imaging camera.
A series of spot of coordinates x = 0, y = 5µm and variable values for the defocus, dz, or of
coordinates y = 5, z = 0µm and variable values for the lateral position dx where generated (the
lateral shift of 5µm allows avoiding the zero order spot).
The corresponding values for the peak intensity as a function of the lateral and axial
displacement are reported in figure 1(c-d).
In diffractive optics, the lateral intensity distribution is determined by the spatially varying
diffraction efficiency, d(dx, dy), defined as the ratio between the integrated intensity, Ispot,
redirected into the desired target spot(s) and the integrated light intensity, I0, of the zero order
spot (SLM off), i.e. d = Ispot/I0. Its value decreases with increasing distance, dx or dy, following
in a good approximation the low numerical aperture approximation (red curve in figure 1 (c)):
2 2
æ sin X ö æ sin Y ö
δ (dx,dy ) = ç ÷ ç ÷ [33] (3)
è X ø è Y ø

6
 æ ö æ ö
where X = ç mpa dx ÷ , Y = ç mpa dy ÷ , m is the refractive index at the sample plane, f is the
ç λf ÷ ç ÷
è ø è λf ø
objective focal plane in water and a is the de-magnified SLM pixel size at the objective’s back
aperture, i.e. a= dSLMf2/f1 (dSLM being the original pixel size which is 20 µm for the model used
in this paper). This expression also sets the qualitative limits for the maximum excitation field:
æ lf1 f ö
2 ´ dxmax = 2 ´ dymax = 2 ´ çç ÷÷ [25, 28].
è m × d SLM f 2 ø

Diffraction efficiency also affects light distribution in the longitudinal direction [25]. The loss
of energy in this case is due to rays coming from pixels imaged at the edge of the pupil. As
pixels diffract light in a limited cone with a half angle of the order of l/2a, where a is the pixel
size, pixels located at the edge of the pupil will diffract on a limited region of the optical axis
mlf
(figure 1(a)), the range of accessible defocus being of the order of 2 ´ dzmax @ 2 ´ ,
2 NA × a
where NA is the numerical aperture of the objective. The more this limit is approached the less
the pixels at the edge contribute to the total intensity at the spot (the corresponding defect of
energy is sent to higher diffraction orders). For the parameters used in our system
2 ´ dxmax = 2 ´ dymax » 180µm and 2 ´ dzmax » 185µm .

To evaluate this effect in our system, we generated a spot of coordinates x=0, y=5µm and
variable values for the defocus, dz, and measured experimentally the corresponding integrated
intensity. The results are reported in figure 1(d) where the green curve is a numerical simulation.
This has been achieved by summing the diffraction patterns from all pixels at the focal sphere.
To take into account that the projection of the pixel at the pupil plane onto the focal sphere gives
rise to a lateral stretch of the original pixel size, a, along x and y directions, the pixel at the focal
sphere has been represented as a rectangle of sides ax, ay. This correction permits a very good
fitting of the experimental data.
A second consequence of the reduced contribution of the side pixels with increasing defocus is
a reduction of the effective numerical aperture and therefore a deterioration of the axial FWHM.
To evaluate this effect in our system, we performed for different values of defocus dz an axial
scan of ± 5 µm. By plotting intensity profiles along x, y direction we determined the lateral
FWHMs by Gaussian fittings. The axial FWHM was measured by fitting the axial intensity by a
Lorentzian provided the diffraction efficiency can be considered as constant on such a small dz

7
range. We obtained experimentally that axial FWHMs varies from 1.2 µm to ~2 µm going from
z = 0 to z = 50 µm and that lateral FWHMs varies from 0.3 µm to ~ 0.4 µm in the same z-range.
These results show that, with the used optical parameters, light efficiency ≥ 50% and an almost
constant lateral and axial FWHMs can be obtained for spot generated in an excitation volume of
80 x 80 x 100 µm3

3. Holographic Photostimulation of the Dendritic Arbor

3.1 Targeting individual dendritic spines
First we wanted to test the resolution and achievable excitation power for photolysis in a
scattering medium and under typical electrophysiological experimental conditions.
To this end, we tested the performance of the system for photostimulation of individual
dendritic spines.
A CA1 hippocampal pyramidal neuron was dialyzed through the patch electrode with Cascade
Blue hydrazide (750µM), a fluorescent dye that can be excited by the 405nm laser beam. Figure
2(a) (left panel) shows the wide-field fluorescence image of a dendritic spine on a dendritic
branch ~ 40µm below the surface. A holographic spot was generated at the position of the
selected spine and the corresponding fluorescence image was acquired (middle panel). The two
images are then superimposed (right panel). The holographically generated spot has a diameter
of ~ 0.6nm (FWHM), which is still close to the value measured at 405nm in a non-scattering
sample (section 2).
To evaluate the ability of the system to generate the necessary light intensity for photolysis a
brief light pulse (0.1ms in duration, 1.5mW at sample plane) was directed at a dendritic spine in
the presence of 2mM of MNI-glutamate (MNI-Glu) (figure 2(b)).
The evoked response is similar in amplitude and kinetics to miniature excitatory postsynaptic
currents, mEPSCs, that have been previously reported for CA1 hippocampal pyramidal neurons
[34] and to light-evoked EPSCs from laser scanning systems [35].
This result suggests that the 3D holographic system has the resolution and usable power to
stimulate individual synaptic spines.

3.2 Effective spatial resolution during photolysis of dendritic branches

In experiments employing photolysis of caged compounds the diffusion of the photoreleased
compound during the time period of illumination also needs to be taken into account when

8
considering the effective spatial resolution. We first determined the lateral resolution by moving
a single excitation spot (6.5mW, 0.5ms) orthogonally across a thin dendrite (1µm in diameter,
located ~ 50µm below the surface of the brain slice) in the presence of 2mM of MNI-Glu, and
recording the corresponding photolysis evoked current response (upper panel, figure 3(a)). The
spot was moved by projecting holograms with increasing lateral shift and keeping the sample at a
fixed position. The FWHM of the peak amplitude as a function of the lateral position was 3.14
µm (lower panel, figure 3(a)). The FWHM for the group data (n=6) was 3.15µm, figure 3(b).
Next, axial resolution was extracted by measuring current elicited while varying the axial
position of the holographic spot (upper panel, figure 3(c)). This was done by projecting
hologram with increasing defocus and keeping the sample at a fixed position. The FWHM
measurement of the current response as a function of axial resolution was 8.7 µm for an
individual dendrite (lower panel, figure 3(c)) and 9.65 µm in group data (n=6) (figure 3(d)).
The observed broadening of FWHM can be accounted by the diffusion of the photoreleased
compound. To roughly simulate this effect we calculated the 3D temporal broadening of the
excitation spot (taking the experimental lateral and axial profiles) due to the diffusion of uncaged
glutamate based on Fick’s second law and the diffusion coefficient D of glutamate in brain tissue
(D= 0.33µm2/ms [36, 37]). We found that a spot of initial lateral FWHM and axial FWHM of 0.3
µm and 1.2 µm, respectively, in the temporal interval of 1ms (3 ms) (current rise time) reaches
corresponding values of 2.1 µm (3.4 µm) and 5.7 µm (9.7 µm). These values are in agreement
with experimental findings, considering the ~1µm diameter of the target dendrite.

3.3 Simultaneous 3D stimulation

In order to demonstrate the 3D capability of digital holography, we carried out focal photolysis
directed at multiple sites located on different axial planes simultaneously. First, we directed
photolysis on two spines on a single dendritic branch but which were located in two different
axial planes separated by 15µm (figure 4(a)). The two dendritic spines were first stimulated
individually and then simultaneously (3.2mW per spot, 0.3ms) in the presence of 2mM of MNI-
Glu. We predicted that if instantaneous 3D photolysis was effective, then the arithmetic sum of
the individual responses would equal that of the response to simultaneous photolysis at
subthreshold amplitudes [9]. Indeed, the spatially summed response of the two axially displaced
spines was the same as the arithmetic sum of the two individual responses (figure 4(a) and 4(b)).

9
We also predicted that attempts to simultaneously stimulate the two spines using photolysis
within a single plane would result in a response that corresponds to that for a single spine.
Indeed, photolysis directed 15µm above the second spine resulted in no observable additional
contribution from the second spine (figure 4(a) and (b)).
We then examined the effectiveness of 3D photolysis using multiple spots on two
dendritic branches located in different axial planes > 20µm apart. The stimulus strength for the
two branches was greater than that employed for dendritic spine stimulation. Photolysis was
directed on 2 locations on the shaft of each dendrite (1.6 mW per spot). The response to
simultaneous photolysis of the two branches was slightly lower than the arithmetic sum of the
responses to stimulating the two branches independently (figure 4(c), individual neuron, figure
4(d), group data). These data again demonstrate the effectiveness of the axial resolution of the
3D stimulation. Photolysis with all the spot held in the same plane (z=0) resulted in no
observable additional response beyond what was expected from the target in that plane. The
small degree of sub-linear summation (10-15%) between branches but not between spines is not
a technical issue of photolysis, but a phenomenon of dendritic excitability (to be discussed in
greater detail in figure 5).

4. Multibranch dendritic integration

One fundamental issue in neuroscience is how inputs arriving on different dendritic branches
are summed. However, it has not been possible to properly address the latter issue using multi-
site photolysis because currently photolysis techniques are limited to 2D light patterns and a
situation where significant portions of multiple dendrites is positioned in a single plane is
unlikely. The only study of integration between two dendritic branches was carried out using two
separate iontophoretic electrodes [38]. However this approach does not provide the full range of
control with respect to the location, timing, and strength of stimulation.
Here we demonstrate that 3D holographic technology overcomes this fundamental limitation
allowing us to conduct integration of multiple dendritic branches.
Figure 5(a) shows a representative set of recording of five different apical oblique dendrites
from acute hippocampal slices. The locations of focal photostimulation (two spots for each
branch, pulse duration 0.3-1ms; total power, 6.5mW) on the five dendrites are schematically
illustrated on the right. The electrical responses to stimulation of the individual branches are

10
shown in the lower traces to the left. And the responses to the simultaneous stimulation of
multiple branches (i.e. spatial integration) are shown in the upper left traces.
It is surprising, on first inspection, that multi-branch summation of the oblique dendrites is sub-
linear whereas summation on the same branch typically is linear to supralinear [9] (Figure 5(a)).
Passive cable theory predicts that for inputs separated by wider distances should sum more
linearly [39]. The observation that multi-branch integration can be significantly sub-linear
suggests that some active mechanisms must be triggered by sub-threshold inputs. A similar sub-
linear summation, albeit less prominent, was observed using iontophretic electrodes to stimulate
two branches [38]. It was suggested that the mechanism for the sub-linear summation in the latter
study was due to the recruitment of the transient IA type potassium channels highly expressed in
the apical dendrites of hippocampal pyramidal neurons [40]. The more robust sub-linear
summation due to summation between larger number of branches could also be eliminated with
the addition of 4-aminopyridine, an antagonist of the IA potassium conductance (unpublished
data).

Summary
In this study we presented an experimental system for the generation of 3D holographic
photostimulation patterns. We demonstrated the flexibility of this approach in different
experimental configurations ranging from the excitation of a single dendritic spine to the
simultaneous stimulation of multiple dendritic branches of hippocampal neurons.
We demonstrated that the use of 3D stimulation to control neuronal excitability permits one to
study spatial integration on a single dendritic branch without the need to keep all of the sites of
photostimulation in the same optical plane. We also demonstrated that 3D holographic light
patterns permit one to investigate how inputs arriving on different dendritic branches are
summed. This is a fundamental question in neuroscience and to date has been inaccessible with
traditional optical methods due to the unlike possibility of multiple interconnected dendrites
forming in a single optical plane.
Other examples for future applications include the study of functional circuitry to map
connectivity of neurons within a 3D space rather than only those within a 2D plane.
The experiments described in this paper have been performed in single photon excitation and
thus where limited to the first 40 µm of the brain slices. It is believed that the use of 3D

11
holography in two photon excitation as proposed by V. Daria et al. [26] will permit to easily
overcome this limit.
3D light patterning can also be achieved with an AODs based set up [41, 42], however, besides
the already mentioned limits in temporal resolution, this approach requires the set up of a bulky
system of difficult alignment. It has significant power losses due to the numerous optical
components. In the case of two photon excitation, requires compensation for chromatic
aberrations and pulse dispersion which are negligible when phase modulation is achieved
through the use of a SLM [43]. Finally the use of a SLM also allows for simultaneous control of
light patterning and correction for optical aberrations, as already suggested for adaptive
correction of optical trap efficiency [44].

Figure legends

Figure 1: 3D holographic light patterns.

(a) Schematic representation explaining how the coordinates at the pupil plane are related to the
spherical wave front in the sample plane.
(b) 3D distribution of spots generated in a cubic pattern. 3D visualization of spots is obtained by
using a double microscope set up where the upper objective generating the holographic light
distribution is scanned (at step of 500 nm) through a fluorescent layer positioned at the focal
plane of the collecting objective. The recorded stack is projected from two perspectives to better
illustrate the point spread function of the individual spots. Upper panel: side view, lower panel:
axial view. 3D image reconstruction has been made with BioImageXD Project, version 0.9.0
beta. Corners of the cube are separated by 20µm.
Lateral and axial FWHM where comprised between 0.3 µm to ~ 0.4 µm and 1.2 µm to ~2 µm,
respectively.
(c) Measured lateral diffraction efficiency, d = Ispot/I0, as a function of the lateral position, dx, of
a spot generated at coordinates y = 5µm and z = 0 and theoretical curve (red line) taking into
account the axial diffraction efficiency of the SLM (equation (3)). (d) Measured axial diffraction
efficiency as a function of the axial position, dz, of a spot generated at coordinates x = 0; y =

12
5µm and theoretical curve (green line) taking into account the axial diffraction efficiency of the
SLM.

Figure 2: Targeting individual dendritic spines.

(a) Left panel: wide field fluorescence image of a dendritic segment loaded with Cascade Blue
hydrazide via the patch electrode. Middle panel: fluorescent image of the laser spot illuminating
the dendritic spine. Right panel: overlay of left and mid panel. (d) Response to stimulation of a
single dendritic spine to a 1.5 mW, 0.1 ms laser pulse in presence of 2mM MNI-glutamate.

Figure 3: Effective spatial resolution during photolysis.

(a) Upper panel: current responses to a constant stimulus (6.5mW, 0.5ms) as the positions of the
stimulation are moved orthogonally across a dendritic branch. Lower panel: Peak current plotted
as a function of lateral position for a single dendrite. (b) Group data as a function of lateral
position. (c) Upper panel: current responses obtained by axially moving the position of the
stimulation through a dendrite. Lower panel: current response for a single dendrite. (d): group
data as a function of axial position.

Figure 4: Simultaneous photostimulation at two different planes.

(a) Two dendritic spines on a steeply inclined dendrite are stimulated separately and then
simultaneously. Spine #2 is located 15µm below spine #1, located at the objective focal plane (0
µm). Simultaneous 3D photolysis of these two spines (3.2mW per spot, 0.3ms) produced linear
summation which equaled the expected arithmetic sum of the individual response to the two
spots. (b) Group data, n=3. Summation across two planes at an average of 19µm axial separation
is linear, whereas stimulation held at a single axial position (0µm) does not add any observable
increased response. (c) The summation of the responses recorded by exciting with two spots (3.2
mW) on two dendrites separated by 20µm in the axial direction is slightly sublinear to the
expected response. (d) Group data, n=4.

Figure 5: Spatial summation on multiple dendritic branches oriented in different directions and
focal planes.

13
(a) Voltage responses of a single CA1 pyramidal neuron to stimulation of individual branches
(lower left) and when multiple branches are stimulated simultaneously (0.3-1ms; total power,
6.5mW) (upper left). The arithmetic sum of the individual branches compared to the actual
response to simultaneous stimulation of multiple branches of this cell is plotted as the red circle
in (b). (b) Group data from seven cells. Sublinear summation is consistently observed up to an
expected depolarization of 30 mV.

Acknowledgements
We thank Roberto Di Leonardo for helpful discussions on the calculation algorithm and
Francesca Anselmi and Aurélien Bègue for the help in setting up the electrophysiological system
for the experiments realized in Paris. EP was supported by a fellowship from the Fondation pour
la Recherche Médicale. CMT acknowledges the financial support from VA Merit Review and
NIH, VE and CMT acknowledge the financial support from Paris school of neuroscience and
Consulate General of France in Los Angeles.

References

1. M. Scanziani and M. Hausser, "Electrophysiology in the age of light," Nature 461, 930-
939 (2009).
2. S. Gasparini and J. C. Magee, "State-dependent dendritic computation in hippocampal
CA1 pyramidal neurons," J Neurosci 26, 2088-2100 (2006).
3. T. Branco, B. A. Clark, and M. Hausser, "Dendritic discrimination of temporal input
sequences in cortical neurons," Science 329, 1671-1675.
4. S. Shoham, D. H. O'Connor, D. V. Sarkisov, and S. S. Wang, "Rapid neurotransmitter
uncaging in spatially defined patterns," Nat Methods 2, 837-843 (2005).
5. B. E. Losavio, V. Iyer, S. Patel, and P. Saggau, "Acousto-optic laser scanning for multi-
site photo-stimulation of single neurons in vitro," J Neural Eng 7, 045002 (2010).
6. M. Matsuzaki, G. C. Ellis-Davies, and H. Kasai, "Three-dimensional mapping of unitary
synaptic connections by two-photon macro photolysis of caged glutamate," J
Neurophysiol 99, 1535-1544 (2008).
7. V. Nikolenko, K. E. Poskanzer, and R. Yuste, "Two-photon photostimulation and
imaging of neural circuits," Nat Methods 4, 943-950 (2007).
8. F. F. Trigo, J. E. Corrie, and D. Ogden, "Laser photolysis of caged compounds at 405
nm: photochemical advantages, localisation, phototoxicity and methods for calibration," J
Neurosci Methods 180, 9-21 (2009).
9. A. Losonczy and J. C. Magee, "Integrative properties of radial oblique dendrites in
hippocampal CA1 pyramidal neurons," Neuron 50, 291-307 (2006).

14
10. P. J. Sjostrom, E. A. Rancz, A. Roth, and M. Hausser, "Dendritic excitability and
synaptic plasticity," Physiol Rev 88, 769-840 (2008).
11. N. Grossman, V. Poher, M. S. Grubb, G. T. Kennedy, K. Nikolic, B. McGovern, R. B.
Palmini, Z. Gong, E. M. Drakakis, M. A. Neil, M. D. Dawson, J. Burrone, and P.
Degenaar, "Multi-site optical excitation using ChR2 and micro-LED array," J Neural Eng
7, 16004 (2010).
12. L. J. Horneck, "128 X 128 Deformable Mirror Device," IEEE Trans. Electron Devices
ED-30, 539 (1983).
13. N. Farah, I. Reutsky, and S. Shoham, "Patterned optical activation of retinal ganglion
cells," Conf Proc IEEE Eng Med Biol Soc 2007, 6369-6371 (2007).
14. S. Wang, S. Szobota, Y. Wang, M. Volgraf, Z. Liu, C. Sun, D. Trauner, E. Y. Isacoff, and
X. Zhang, "All optical interface for parallel, remote, and spatiotemporal control of
neuronal activity," Nano Lett 7, 3859-3863 (2007).
15. Z. V. Guo, A. C. Hart, and S. Ramanathan, "Optical interrogation of neural circuits in
Caenorhabditis elegans," Nat Methods 6, 891-896 (2009).
16. C. W. Liang, M. Mohammadi, M. D. Santos, and C.-M. Tang, "Patterned
photostimulation with digital micromirror devices to investigate dendritic integration
across branch points," J of Visualized Experiments in press(2011).
17. J. E. Curtis, B. A. Koss, and D. G. Grier, "Dynamic holographic optical tweezers.," Opt.
Communi. 207, 169 (2002).
18. H. Melville, G. F. Milne, G. C. Spalding, W. Sibbett, K. K. Dholakia, and D. D.
McGloin, "Optical trapping of three dimensional structures using dynamic holograms.,"
Opt. Express 11, 3562 (2003).
19. J. Leach, G. Sinclair, P. Jordan, J. Courtial, M. J. M. J. Padgett, J. Cooper, and Z. J.
Laczik, " 3D manipulation of particles into crystal structures using holographic optical
tweezers," Opt. Express 12, 220 (2004).
20. V. Emiliani, D. Cojoc, E. Ferrari, V. Garbin, C. Durieux, M. Coppey, and E. Di Fabrizio,
"Wave front engineering for living cells microscopy," Optics Express 13, 1395-1405
(2005).
21. P. J. Rodrigo, V. R. Daria, and J. Glückstad, "Four-dimensional optical manipulation of
colloidal particles," Applied Physics Letters 86, 074103 (2005).
22. E. Papagiakoumou, F. Anselmi, A. Begue, V. de Sars, J. Gluckstad, E. Y. Isacoff, and V.
Emiliani, "Scanless two-photon excitation of channelrhodopsin-2," Nat Methods 7, 848-
854 (2010).
23. W. Gobel, B. M. Kampa, and F. Helmchen, "Imaging cellular network dynamics in three
dimensions using fast 3D laser scanning," Nat Methods 4, 73-79 (2007).
24. C. Lutz, T. S. Otis, V. de Sars, S. Charpak, D. A. DiGregorio, and V. Emiliani,
"Holographic photolysis of caged neurotransmitters," Nat Methods 5, 821-827 (2008).
25. L. Golan, I. Reutsky, N. Farah, and S. Shoham, "Design and characteristics of
holographic neural photo-stimulation systems," J Neural Eng 6, 66004 (2009).
26. V. R. Daria, C. Stricker, R. Bowman, S. Redman, and H. A. Bachor, "Arbitrary multisite
two-photon excitation in four dimensions," Applied Physics Letters 95, - (2009).
27. V. Nikolenko, B. O. Watson, R. Araya, A. Woodruff, D. S. Peterka, and R. Yuste, "SLM
Microscopy: Scanless Two-Photon Imaging and Photostimulation with Spatial Light
Modulators," Front Neural Circuits 2, 5 (2008).
28. M. Zahid, M. Velez-Fort, E. Papagiakoumou, C. Ventalon, M. C. Angulo, and V.
Emiliani, "Holographic photolysis for multiple cell stimulation in mouse hippocampal
slices," PLoS One 5, e9431 (2010).

15
29. B. Richards and E. Wolf, "Electromagnetic diffraction in optical systems. II. Structure of
the image field in an aplanatic system," Proc. Roy. Soc. Lond. Ser. A 253, 358-379
(1959).
30. J. Leach, K. Wulff, G. Sinclair, P. Jordan, J. Courtial, L. Thomson, G. Gibson, K.
Karunwi, J. Cooper, Z. J. Laczik, and M. J. M. J. Padgett, "Interactive approach to optical
tweezers control," Appl Opt 45, 897 (2006).
31. E. J. Botcherby, R. Juskaitis, M. J. Booth, and T. Wilson, "Aberration-free optical
refocusing in high numerical aperture microscopy," Optics Letters 32, 2007-2009 (2007).
32. R. Di Leonardo, F. Ianni, and G. Ruocco, "Computer generation of optimal holograms for
optical trap arrays," Optics Express 15, 1913-1922 (2007).
33. M. Born and E. Wolf, Principles of optics : electromagnetic theory of propagation,
interference and diffraction of light, 7th ed. ed. (Cambridge University Press, Cambridge,
1999), p. 1200p.
34. R. C. Carroll, R. A. Nicoll, and R. C. Malenka, "Effect of PKA and PKC on miniature
excitatory postsynaptic currents in CA1 pyramidal cells," J Neurophysiol 80, 2797
(1998).
35. A. A. Bagal, J. P. Kao, C. M. Tang, and S. M. Thompson, "Long-term potentiation of
exogenous glutamate responses at single dendritic spines," Proc Natl Acad Sci U S A
102, 14434-14439 (2005).
36. T. S. Otis, Y. C. Wu, and L. O. Trussell, "Delayed clearance of transmitter and the role of
glutamate transporters at synapses with multiple release sites," J Neurosci 16, 1634-1644.
(1996).
37. D. A. DiGregorio, J. S. Rothman, T. A. Nielsen, and R. A. Silver, "Desensitization
properties of AMPA receptors at the cerebellar mossy fiber granule cell synapse.," J.
Neurosci. 27, 8344–8357 (2007).
38. S. Cash and R. Yuste, "Input summation by cultured pyramidal neurons is linear and
position-independent," J Neurosci 18, 10-15 (1998).
39. W. Rall, "Distinguishing theoretical synaptic potentials computed for different soma-
dendritic distributions of synaptic input," J Neurophysiol 30, 1138-1168 (1967).
40. D. A. Hoffman, J. C. Magee, C. M. Colbert, and D. Johnston, "K+ channel regulation of
signal propagation in dendrites of hippocampal pyramidal neurons," Nature 387, 869-875
(1997).
41. G. D. Reddy and P. Saggau, "Fast three-dimensional laser scanning scheme using
acousto-optic deflectors," J Biomed Opt 10, 064038 (2005).
42. P. A. Kirkby, K. M. Srinivas Nadella, and R. A. Silver, "A compact Acousto-Optic Lens
for 2D and 3D femtosecond based 2-photon microscopy," Opt Express 18, 13721-13745
(2010).
43. E. Papagiakoumou, V. de Sars, D. Oron, and V. Emiliani, "Patterned two-photon
illumination by spatiotemporal shaping of ultrashort pulses," Optics Express 16, 22039-
22047 (2008).
44. K. D. Wulff, D. G. Cole, L. C. Robert, R. DiLeonardo, L. Jonathan, J. Cooper, G.
Graham, and M. J. Padgett, " Aberration correction in holographic optical tweezers," Opt.
Express 14, 4170 (2006).

16
(a) (b)
Focal sphere
P (xp, yp, zp)

F (0, 0, f) 10 µm
z
z x
dz
N (xn, yn, zn)
Sample volume
y y
x z
Pupil plane

Normalized diffraction efficeincy, d(dz)

Normalized diffraction efficeincy, d(dx)

(c) 1.0
(d) 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
-60 -40 -20 0 20 40 60 -100 -50 0 50 100
dx (µm) dz (µm)

Figure 1
(a) (b)
Cascade blue- Holographic Overlay
labeled dendrite excitation spot

Figure 2
(a) (c)

(b) (d)

Lateral position (µm) Axial position (µm)

Figure 3
(a) (c)

(d)
(b)

Figure 4
(a)

(b)

Figure 5